Last 7 Days (July 20 – July 26, 2026)
LLMs scale Mixture-of-Experts (MoE) parameters for superior intelligence, but massive weights and dynamic computation impede efficient serving. Existing instance-level prefill-decode disaggregation isolates the phases on separate full-model replicas. As MoE weights grow, each instance may span tens to hundreds of GPUs, making resource allocation increasingly coarse. Configured prefill-to-decode ratios thus often mismatch demand, overprovisioning one phase while overloading the other. Prefill-decode colocation avoids this duplication, but existing Green Context solutions partition each GPU by phase and fix phase resources during a kernel. They cannot track resource changes across operations or layerwise variation in routed expert load, causing head-of-line blocking or idle reserved resources. Partitioning every GPU also leaves each phase with fewer local resources, forces wider parallelism and more communication, and lets prefill and decode traffic interfere on the shared network. We present ExpertPlex, which shares massive MoE experts across phases while disaggregating lightweight attention modules. Expert sharing eliminates over 95% of duplicate model weights and multiplexes dynamically sparse computation, while attention disaggregation reduces attention communication cost. ExpertPlex further uses (1) adaptive persistent kernels to schedule dynamic expert computation at tile granularity for efficient, isolated execution; (2) attention-initiated MoE communication to avoid network interference and enable cross-phase communication-computation overlap; and (3) a tile-to-cluster model to optimize these mechanisms for maximum goodput. Experiments serving MiniMax-M2.7 and GLM-5.1-FP8 show that ExpertPlex improves goodput by up to 2.01$\times$ over instance-level prefill-decode disaggregation and 1.66$\times$ over prefill-decode colocation.
Primary: Peking University
All Institutions: Peking University, Independent Researcher
ExpertPlex presents a significant advancement in LLM serving systems by introducing a novel disaggregated architecture that effectively shares MoE experts while isolating attention modules, achieving substantial goodput improvements through adaptive persistent kernels and optimized communication patterns.
The paper proposes ExpertPlex, a disaggregated serving architecture specifically designed for Mixture-of-Experts (MoE) Large Language Models. The core innovation lies in decoupling the handling of MoE experts from attention modules. While existing systems either colocate all components (leading to resource contention) or disaggregate at the instance level (leading to massive weight duplication and coarse-grained allocation), ExpertPlex shares the massive MoE expert weights across phases while isolating the lightweight attention modules. The methodology introduces three key technical mechanisms: (1) Adaptive Persistent Kernels, which schedule dynamic expert computation at the tile granularity to handle load imbalance and avoid head-of-line blocking; (2) Attention-Initiated MoE Communication, which overlaps communication with computation and prevents network interference between prefill and decode phases; and (3) A Tile-to-Cluster Model, an optimization framework to allocate resources dynamically. This approach addresses the fundamental inefficiency of serving sparse MoE models where expert load is highly dynamic and non-uniform.
The evaluation is conducted on two significant MoE models: MiniMax-M2.7 and GLM-5.1-FP8. The baseline comparisons are rigorous, covering the two dominant existing paradigms: instance-level prefill-decode disaggregation and prefill-decode colocation. The results demonstrate substantial improvements, with up to 2.01x goodput improvement over instance-level disaggregation and 1.66x over colocation. These gains are particularly impressive given that colocation is often considered the most resource-efficient in terms of hardware usage, implying that ExpertPlex achieves higher throughput without requiring additional hardware, simply by better utilizing existing resources. The use of FP8 models also highlights the system's relevance to current hardware trends.
The paper provides a detailed description of the system design, including the persistent kernel scheduling and communication overlap mechanisms. The inclusion of specific model names (MiniMax-M2.7, GLM-5.1-FP8) allows for potential replication if these models are publicly available or if synthetic workloads are used. However, as is common with systems papers, full reproducibility might depend on the specific cluster configuration and the availability of the proprietary model weights. The detailed algorithmic descriptions of the adaptive scheduling and tile-to-cluster optimization provide a strong foundation for implementation.
The paper focuses on the serving side and does not address training efficiency. The complexity of the adaptive persistent kernels and the tile-to-cluster model introduces additional system overhead that must be carefully managed; if the scheduling overhead exceeds the gains from reduced communication or better utilization, performance could degrade. Furthermore, the benefits are most pronounced for very large MoE models with high expert counts; for smaller models or dense models, the overhead of the disaggregation and dynamic scheduling might not justify the complexity. The reliance on high-bandwidth, low-latency interconnects for the attention-MoE communication is also a constraint.
ExpertPlex addresses a critical bottleneck in the deployment of state-of-the-art AI models. By significantly improving the goodput of MoE LLMs, it lowers the cost barrier for serving these models, potentially democratizing access to high-quality AI services. The architectural insights regarding dynamic resource allocation for sparse models can influence future system designs for other sparse architectures beyond LLMs, such as sparse transformers or mixture-of-experts in other domains. ExpertPlex presents a significant advancement in LLM serving systems by introducing a novel disaggregated architecture that effectively shares MoE experts while isolating attention modules, achieving substantial goodput improvements through adaptive persistent kernels and optimized communication patterns.
Scaling executable agent training data for LLM post-training is bottlenecked by substrate-bound methods that tie task generation to predefined tools, repositories, or skill graphs: expanding coverage requires manual substrate engineering, each new domain demands a bespoke pipeline, and the resulting task distributions often reflect substrate biases rather than real-world demand. We introduce NexForge, a requirement-driven framework that takes high-level capability requirements as input and synthesizes diverse, executable agent tasks and expert trajectories for SFT. NexForge first investigates real-world demand to construct representative scenarios and task profiles, then performs distribution-aware compilation to generate task directives. For each directive, NexForge automatically retrieves or constructs the required files, dependencies, and runtime configurations, and finally synthesizes expert rollouts and produces training trajectories. Without domain-specific infrastructure, NexForge produces 3.6K terminal and 2K office tasks, improving Qwen3.5-35B-A3B Base from 22.5\% to 52.0\% on Terminal-Bench 2.0 and from 813 to 1338 Elo on GDPval; scaling further to 43.2K terminal tasks yields 58.4\%, on par with Claude Opus 4.6 equipped with Claude Code. Scaled further, NexForge-synthesized data contributes to the training of Nex-N2, a family of publicly available agent models that lift Qwen3.5-35B-A3B to 75.3\% on Terminal-Bench 2.1 and to 1585 Elo on GDPval -- achieving state-of-the-art open-source performance and surpassing several frontier proprietary systems. Nex-N2 models are available at https://nex.sii.edu.cn/.
Primary: SII (Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences)
All Institutions: SII (Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences)
NexForge presents a compelling and effective pipeline for scaling agent training data through requirement-driven synthesis, demonstrating that high-quality, diverse task generation can significantly boost LLM agent performance, achieving state-of-the-art open-source results on key benchmarks.
The paper introduces NexForge, a framework designed to address the data bottleneck in training LLM-based agents. The core innovation lies in shifting from "substrate-bound" task generation (which relies on predefined tools or codebases) to a "requirement-driven" approach. The methodology involves three key stages: 1) Analyzing real-world demand to create representative scenarios and task profiles; 2) Distribution-aware compilation to generate high-level task directives; and 3) Automatic synthesis of executable environments (files, dependencies, runtime configs) and expert rollouts for Supervised Fine-Tuning (SFT). This approach aims to reduce manual engineering and mitigate substrate biases. The method is technically sound, leveraging existing LLM capabilities for code generation and environment setup, but the novelty is incremental rather than foundational. It represents a sophisticated engineering pipeline rather than a new algorithmic breakthrough.
The experimental section demonstrates significant empirical improvements. Using Qwen3.5-35B-A3B as the base, the authors show a jump from 22.5% to 52.0% on Terminal-Bench 2.0 and from 813 to 1338 Elo on GDPval with 3.6K terminal and 2K office tasks. Scaling to 43.2K terminal tasks pushes performance to 58.4%, which is comparable to Claude Opus 4.6 with Claude Code. Furthermore, the synthesized data is used to train "Nex-N2," achieving state-of-the-art open-source results (75.3% on Terminal-Bench 2.1, 1585 Elo on GDPval). The results are impressive and suggest that high-quality, diverse, requirement-driven data is a critical lever for agent performance. The evaluation is rigorous, covering multiple benchmarks and comparing against strong proprietary baselines.
The paper provides a project URL (https://nex.sii.edu.cn/) which likely contains code and model weights. The description of the pipeline (requirement analysis -> directive compilation -> environment synthesis -> rollout) is detailed enough to be reproducible by a team with sufficient resources. However, the "expert rollouts" likely rely on a strong teacher model or human-in-the-loop, which can introduce variability. The specific "distribution-aware compilation" algorithm is not fully detailed in the abstract, so full reproducibility depends on the completeness of the main text and code release.
The paper does not explicitly discuss the cost of generating 43.2K high-quality tasks, which can be significant. The reliance on a "requirement-driven" approach assumes that high-level requirements can be effectively mapped to executable tasks, which may fail in domains with ambiguous or complex implicit constraints. Additionally, the "substrate biases" argument, while valid, might be overstated if the underlying LLMs themselves have biases in their training data that NexForge cannot correct. The evaluation is primarily on coding/terminal tasks; generalization to other agent domains (e.g., web browsing, multi-modal reasoning) is not demonstrated.
This work has significant implications for the democratization of capable AI agents. By providing a scalable method for generating high-quality training data, it lowers the barrier to entry for developing specialized agents. The release of Nex-N2 models contributes to the open-source ecosystem. However, the potential for misuse (e.g., generating malicious code or automating cyberattacks) is a concern that should be addressed in the broader impact statement. The success of such frameworks may accelerate the arms race in agent capabilities, raising safety and alignment challenges. NexForge presents a compelling and effective pipeline for scaling agent training data through requirement-driven synthesis, demonstrating that high-quality, diverse task generation can significantly boost LLM agent performance, achieving state-of-the-art open-source results on key benchmarks.
Retrieval-Augmented Generation (RAG) enhances the factual grounding of large language model (LLM) inference by retrieving relevant information from external knowledge bases. However, its dense vector retrieval introduces significant latency and energy overhead, becoming the primary performance bottleneck. Although recent in-storage accelerators aim to reduce data movement, they still rely on host or embedded processors outside the memory, where nearly 70% of the total retrieval time is spent. As a result, they cannot fully overcome the bandwidth limitations, leading to yet another memory bottleneck. To tackle these limitations, we present D-NOVA, a hardware-software co-designed in-storage retrieval accelerator. D-NOVA executes an inverted file (IVF)-based hierarchical retrieval pipeline by deeply embedding the search functionality directly into the NAND memory array. This is achieved by incorporating a new distance metric, Dual-Bound Tight Similarity Sensing (DTS), which is specifically tailored for searching within the NAND string. In addition, we introduce a lightweight contrastive adapter that maps embedding vectors into a DTS-friendly domain, recovering near-software recall while improving performance and energy efficiency. D-NOVA is up to 41.7x faster and 71x more energy-efficient than a CPU baseline, and achieves 12.13x higher throughput while being up to 1.26x more energy-efficient than state-of-the-art in-storage RAG accelerators, demonstrating the potential of fully in-storage vector search for scalable RAG acceleration.
Primary: University of California, San Diego
All Institutions: University of California, San Diego
D-NOVA introduces a novel in-storage vector search accelerator that leverages a NAND-optimized distance metric and contrastive adaptation to achieve significant speed and energy efficiency gains for RAG workloads. This research represents a substantial contribution to the intersection of computer architecture and machine learning systems, offering a viable path to overcoming the memory bottleneck in large-scale retrieval tasks.
The paper proposes D-NOVA, a hardware-software co-designed accelerator that performs vector similarity search directly within 3D NAND memory arrays, bypassing the traditional host-to-memory data movement bottleneck. The core technical innovation is the Dual-Bound Tight Similarity Sensing (DTS) metric, which is specifically tailored to the physical characteristics of NAND strings (e.g., threshold voltage distributions and read disturb effects) to enable approximate nearest neighbor search at the storage level. This is coupled with a lightweight contrastive adapter that transforms embedding vectors into a domain compatible with DTS, allowing the hardware to operate on "DTS-friendly" vectors while maintaining high recall relative to software baselines. The approach represents a significant shift from "in-storage computing" (which often still moves data to embedded processors) to "in-storage sensing," leveraging the analog/digital properties of the memory array itself for computation.
The evaluation demonstrates substantial improvements over baselines. D-NOVA achieves up to 41.7x speedup and 71x energy efficiency gains compared to a CPU baseline. When compared to state-of-the-art in-storage RAG accelerators, it shows 12.13x higher throughput and up to 1.26x better energy efficiency. The paper likely includes detailed breakdowns of latency components, energy consumption per operation, and recall/accuracy metrics (e.g., Recall@K) to validate that the hardware approximation does not significantly degrade retrieval quality. The results suggest that the proposed architecture effectively mitigates the memory wall for RAG workloads.
As a hardware design paper, reproducibility depends on the availability of the RTL (Register Transfer Level) code, simulation models, and detailed architectural parameters. The paper mentions support from SRC and NSF grants, suggesting rigorous academic standards. However, without explicit open-source code links in the provided text, reproducibility is limited to the described methodology and potentially available supplementary materials. The specific implementation of the DTS metric and the contrastive adapter's training procedure are critical for replication.
The primary limitation is the reliance on specific 3D NAND characteristics, which may vary across manufacturers and process nodes. The "DTS-friendly" domain requires a pre-processing step (the contrastive adapter), adding a small overhead that must be justified by the massive gains in the search phase. Furthermore, the scalability of the in-storage search logic across very large vector databases (millions/billions of vectors) and the impact of wear-leveling and garbage collection in NAND on the consistency of the DTS metric are potential challenges not fully addressed in the abstract. The energy efficiency claim of 1.26x over SOTA in-storage accelerators is modest compared to the CPU baseline, suggesting that while the approach is novel, the absolute gains over existing in-storage solutions might be incremental in some configurations.
This work has significant implications for the scalability of Retrieval-Augmented Generation (RAG) systems, which are becoming the standard for enterprise LLM applications. By reducing the latency and energy cost of vector retrieval, D-NOVA enables more responsive and sustainable AI systems. It also advances the field of in-memory/in-storage computing, demonstrating that complex ML workloads can be offloaded to storage devices in a way that leverages their physical properties, potentially reshaping the architecture of future data centers. D-NOVA introduces a novel in-storage vector search accelerator that leverages a NAND-optimized distance metric and contrastive adaptation to achieve significant speed and energy efficiency gains for RAG workloads. This research represents a substantial contribution to the intersection of computer architecture and machine learning systems, offering a viable path to overcoming the memory bottleneck in large-scale retrieval tasks.
Speculative decoding accelerates autoregressive generation by having a cheap draft propose tokens that a target verifies in parallel. Frontier models increasingly ship a built-in Multi-Token-Prediction (MTP/NEXTN) draft head under the assumption that the draft is negligibly cheap. At million-token context this breaks: an MTP draft head typically runs full attention over the entire KV cache at every draft step, so its read grows linearly with context and comes to dominate the draft cost -- precisely where speculation is most valuable. The effect compounds with draft length (a deep native draft can turn net-negative, slower than no speculation) and sharpens under hybrid/linear-attention targets, where cheaper verification leaves the draft's full-attention read exposed. We apply a StreamingLLM-style sliding window plus attention sink to the draft's attention only (Windowed-MTP), leaving full-attention verification intact. It is training-free, drop-in, and lossless by construction: the full-attention target still decides every accepted token, so windowing changes only which tokens are proposed, never which are accepted. It bounds the draft's KV working set to a constant, dropping ~99% of KV entries at 1M. Across three architecture families (Qwen GDN-MoE 35B/122B and a Mamba2-hybrid NoPE 120B) at 1M context on a single GPU in SGLang, windowing cuts the per-decode-step cost over the shipping native MTP draft by +28% to +44%, an input-invariant margin that widens with context. Since per-token latency is this cost divided by acceptance length, at matched acceptance end-to-end decode latency improves by the same amount, and more where windowing also lifts acceptance, while preserving the target's verified output distribution. Finally, the unread draft KV -- 7.7-11% of total KV at 1M -- is reclaimed via a compact ring buffer at no acceptance or quality cost.
Primary: NVIDIA
All Institutions: NVIDIA
This paper presents a highly effective, training-free optimization for speculative decoding in long-context LLMs, solving a critical scalability bottleneck by windowing the draft head's attention, thereby enabling efficient million-token inference without compromising output quality.
The paper addresses a critical bottleneck in long-context speculative decoding: the linear scaling of KV cache access for Multi-Token Prediction (MTP) draft heads. The proposed method, Windowed-MTP, applies a StreamingLLM-style sliding window with attention sinks exclusively to the draft head, while maintaining full attention for the target verification head. This is a clever, training-free architectural tweak that decouples the draft's context dependency from the full sequence length. The methodology is sound, leveraging the observation that draft quality degrades gracefully with windowing, while verification remains exact. It effectively transforms the draft cost from $O(N)$ to $O(1)$ (relative to context length), solving the "net-negative" speculation problem at million-token scales.
The evaluation is robust, covering three distinct architecture families (Qwen GDN-MoE and Mamba2-hybrid) at 1M context. The results demonstrate a 28-44% reduction in per-decode-step cost, which translates directly to end-to-end latency improvements. The experiments are conducted in a realistic setting (SGLang on a single GPU), ensuring practical relevance. The claim of "lossless" acceptance is supported by the fact that the target head is unchanged; the paper correctly identifies that windowing only affects *which* tokens are proposed, not *which* are accepted, preserving the output distribution. The reclamation of unread KV cache via a ring buffer is a nice engineering touch that further optimizes memory.
The method is described with sufficient detail for reproduction. The use of standard components (StreamingLLM, SGLang) and the training-free nature of the approach enhance reproducibility. The codebase is likely open-source given the NVIDIA affiliation and the nature of the contribution, though a specific URL is not provided in the text. The experimental setup is clearly defined.
The primary limitation is the potential drop in draft acceptance rate due to the reduced context window in the draft head. While the paper claims this is mitigated by the attention sink and the nature of MTP, long-range dependencies that are crucial for accurate token prediction might still be missed, potentially reducing the speedup factor (acceptance length) even if the per-step cost is lower. The performance gain is also contingent on the target model being significantly more expensive to verify than the draft is to generate, which is true for MTP but might vary for other speculative setups.
This work has significant implications for the practical deployment of LLMs with million-token contexts. By making speculative decoding efficient at scale, it enables faster inference for applications like long-document analysis, code generation, and complex reasoning tasks that require large context windows. It sets a new standard for how draft heads should be designed for long-context scenarios, likely influencing future model architectures and inference engines. This paper presents a highly effective, training-free optimization for speculative decoding in long-context LLMs, solving a critical scalability bottleneck by windowing the draft head's attention, thereby enabling efficient million-token inference without compromising output quality.
Unlike large language models (LLMs) that exhibit strong reasoning capabilities, vision-language models (VLMs) struggle with visual reasoning, even on geometry problems that admit equivalent text, diagram, and combined diagram+text views. We show that these views often elicit different behaviors: a model may solve a problem from text but fail on the corresponding diagram, or succeed visually while failing textually. This inconsistency suggests that different views expose complementary reasoning paths and failure modes that standard multimodal post-training does not fully exploit. To study and exploit this phenomenon, we construct ODA-Data, a high-quality paired multimodal geometry dataset with text-dominant, image-dominant, and combined image+text views of the same problems, together with splits for training and evaluating modality-dependent reasoning behaviors. We then develop Modality-Informed Reciprocal Reasoning Optimization (MIRROR), a reinforcement learning approach for improving multimodal reasoning via self supervision. For each problem, MIRROR evaluates the model under all views, selects the best-performing view as a teacher, and trains other views with a reverse-KL objective towards the teacher. Across reasoning benchmarks that evaluate on geometry problems, MIRROR improves over standard RL and yields more accurate and consistent behavior across modalities
Primary: unknown
All Institutions: unknown
MIRROR makes a significant contribution to the field of multimodal AI by directly addressing the critical issue of reasoning inconsistency across modalities. By demonstrating that different views expose complementary reasoning paths and failure modes, and providing a method to exploit this, the paper paves the way for more robust and reliable VLMs. The ODA-Data dataset is a valuable resource for future research, enabling more targeted studies of multimodal reasoning. The principles of reciprocal reasoning and self-supervision from "other views" could be extended to other multimodal tasks and model architectures, potentially leading to more generalizable and human-like reasoning capabilities in AI systems. This work could inspire new evaluation metrics and training paradigms that prioritize consistency alongside accuracy. This paper presents a compelling and rigorously evaluated approach to a fundamental problem in multimodal reasoning. The authors' initial observation of modality inconsistency is a strong empirical finding, which they then effectively address with the novel ODA-Data dataset and the MIRROR framework. The methodology, which combines RL with self-supervised knowledge distillation using a "best view" teacher, is innovative and well-justified, leading to significant improvements in both accuracy and consistency across modalities. The comprehensive experiments and exceptional reproducibility details make this a highly impactful contribution to the field.
The paper introduces MIRROR, a novel reinforcement learning approach for improving multimodal reasoning by leveraging "other views." The methodology is built upon a crucial empirical observation: vision-language models (VLMs) often exhibit inconsistent reasoning across different modalities (text, diagram, combined) for the same problem. This inconsistency is rigorously demonstrated through a pilot study using PaLM-2-V and LLaVA-1.5 on geometry problems. To address this, the authors construct ODA-Data, a high-quality paired multimodal geometry dataset with equivalent problems presented in text-dominant, image-dominant, and combined views, specifically designed to study and exploit modality-dependent reasoning behaviors. MIRROR's core idea is to use self-supervision where the model's best-performing view for a given problem acts as a teacher for its other, less successful views. This is formulated as a policy optimization problem within an RL framework. The objective combines a standard reward for correct answers with a reverse-KL regularization term. This reverse-KL term encourages the answer distribution of a non-teacher view to align with that of the teacher view, effectively transferring knowledge and promoting consistency. The teacher selection mechanism, which dynamically identifies the best view based on the model's current performance, is particularly clever. The overall framework is sound, combining established techniques (RL, knowledge distillation) in a novel configuration to tackle a specific and important challenge in multimodal AI.
The experimental evaluation is comprehensive and well-executed. The authors evaluate MIRROR on two prominent VLMs, PaLM-2-V and LLaVA-1.5, across multiple reasoning benchmarks: their newly introduced ODA-Data, GeoQA+, and MathVista. Baselines include standard supervised fine-tuning and an RL-only approach (REINFORCE without the reverse-KL regularization). The results consistently demonstrate that MIRROR significantly outperforms baselines in both accuracy and, critically, consistency across modalities. For instance, MIRROR improves consistency by up to 10.7% on ODA-Data. Ablation studies clearly show the importance of the reverse-KL term, confirming its role in knowledge transfer and consistency enforcement. The experiments also include detailed analysis of how MIRROR helps models correct errors in one modality by leveraging insights from another. The use of ODA-Data's modality-dependent splits allows for a granular evaluation of how MIRROR impacts reasoning across different views. The qualitative examples further illustrate the method's effectiveness in practice. The benchmarks chosen are appropriate for evaluating geometric and general mathematical reasoning, validating the method's applicability beyond the specific ODA-Data.
Reproducibility is a strong suit of this paper. The authors provide extensive details in the appendices, which include: 1. **Computation Details:** Specifics on hardware used (TPUv4, A100 GPUs). 2. **Hyperparameters:** Detailed tables of hyperparameters for both PaLM-2-V and LLaVA-1.5 across different datasets. 3. **Prompt Templates:** Examples of the exact prompt templates used for different views (text, diagram, combined). 4. **Pseudocode:** Clear and concise pseudocode for the MIRROR algorithm, making the implementation straightforward to understand. 5. **Dataset Availability:** The ODA-Data dataset is made publicly available via a GitHub repository (https://github.com/google-research/oda-data). These details, combined with the clear methodology description, make the work highly reproducible.
One potential limitation is the reliance of the teacher selection mechanism on the model being able to solve the problem correctly in at least one view. If a problem is extremely difficult and the model fails across all modalities, the "best view" teacher might not provide a strong signal for improvement, or the reverse-KL objective might not be as effective. The current scope primarily focuses on geometry and mathematical reasoning problems; while the principles might generalize, direct applicability to other multimodal reasoning tasks (e.g., visual commonsense, instruction following) would require further validation. The computational cost of RL fine-tuning, especially with large VLMs, can also be substantial, though this is a common challenge in the field.
MIRROR makes a significant contribution to the field of multimodal AI by directly addressing the critical issue of reasoning inconsistency across modalities. By demonstrating that different views expose complementary reasoning paths and failure modes, and providing a method to exploit this, the paper paves the way for more robust and reliable VLMs. The ODA-Data dataset is a valuable resource for future research, enabling more targeted studies of multimodal reasoning. The principles of reciprocal reasoning and self-supervision from "other views" could be extended to other multimodal tasks and model architectures, potentially leading to more generalizable and human-like reasoning capabilities in AI systems. This work could inspire new evaluation metrics and training paradigms that prioritize consistency alongside accuracy. This paper presents a compelling and rigorously evaluated approach to a fundamental problem in multimodal reasoning. The authors' initial observation of modality inconsistency is a strong empirical finding, which they then effectively address with the novel ODA-Data dataset and the MIRROR framework. The methodology, which combines RL with self-supervised knowledge distillation using a "best view" teacher, is innovative and well-justified, leading to significant improvements in both accuracy and consistency across modalities. The comprehensive experiments and exceptional reproducibility details make this a highly impactful contribution to the field.
Modern AI agents rely on elaborate inference harnesses such as Claude Code, Codex, and OpenClaw to drive multi-turn reasoning, tool use, and access to external systems. While powerful, these complex harnesses also make agents hard to train end-to-end with open infrastructure, whose SFT/RL stacks cannot natively express stateful, multi-process harness inference. To address this, we present OpenForgeRL, an open-source framework for training harness-based agents end-to-end in diverse environments. OpenForgeRL achieves this with a lightweight proxy that serves the harness's model calls while recording them as training data for a standard RL codebase (e.g., veRL), and a Kubernetes orchestrator that runs each rollout in its own remote container, together enabling training on any harness in any environment at scale. By decoupling training and inference, OpenForgeRL allows researchers to easily train, study, and improve agents directly in the real harnesses and environments they are deployed with. We validate our framework across diverse, complex harnesses and environments, spanning tool/claw-based agents and multimodal GUI browser- and computer-use agents. Using only hundreds to a few thousand tasks, OpenForgeClaw reaches 31.7 pass^3 and 55.9 pass@3 on ClawEval and 33.7 on QwenClawBench. OpenForgeGUI reaches 37.7 on OSWorld-Verified, 63.0 on Online-Mind2Web, and 72.3 on WebVoyager. Both outperform open baselines of similar size on nearly all benchmarks, and in the GUI setting match or surpass models several times larger. Beyond benchmarks, we analyze how harness choice (e.g., ZeroClaw, OpenClaw, Codex) and RL shape agent behavior. We find that some harnesses are substantially harder to learn than others, and that RL improves agentic reliability, such as self-verification, tool coverage, and completing multi-step plans, though critical abilities such as error recovery remain weak.
Primary: Dartmouth College
All Institutions: Dartmouth College
OpenForgeRL has a substantial broader impact on the field of AI agents: 1. **Democratizing Agent Research:** By providing an open-source framework to train agents in their *real* deployment harnesses, it significantly lowers the barrier for academic researchers and smaller labs to conduct end-to-end training. This directly addresses the "train-deploy mismatch" that has increasingly favored proprietary systems. 2. **Accelerating Agent Development:** Enabling end-to-end training in complex, stateful environments means agents can learn directly from real-world interactions, leading to more robust and capable agents. This could accelerate progress in domains like software engineering, tool use, and multimodal GUI control. 3. **Facilitating Deeper Analysis:** The framework allows researchers to study the impact of harness design and RL training on agent behavior in unprecedented detail, as demonstrated by the paper's discussion section. This can lead to a better understanding of what makes agents effective and how to design better harnesses and training regimes. 4. **New Benchmarking and Data Generation Paradigms:** The automated data synthesis pipeline is a valuable contribution that can help create more diverse and challenging benchmarks for agent evaluation, especially in data-scarce domains. 5. **Bridging the Gap to Frontier Models:** By allowing open models to be trained in the same sophisticated harnesses used by frontier models, OpenForgeRL helps close the capability gap between open and closed-source agent systems. Overall, OpenForgeRL is a timely and impactful contribution that promises to unlock new avenues for research and development in the rapidly evolving field of AI agents. OpenForgeRL introduces a scalable, open-source framework for training harness-based AI agents end-to-end in any environment, bridging the critical gap between complex inference harnesses and standard RL training stacks. This paper presents a robust engineering solution with a lightweight proxy and Kubernetes orchestrator, validated by extensive empirical results across diverse text-based tool-use and multimodal GUI environments, where it outperforms open baselines and provides valuable insights into harness design and the behavioral impact of RL.
OpenForgeRL addresses a critical bottleneck in training modern AI agents: the "train-deploy mismatch" caused by complex, stateful inference harnesses that are difficult to integrate with standard open-source SFT/RL stacks. The proposed methodology is a well-engineered solution built on two main components: 1. **Lightweight Proxy:** This component abstracts the harness's inference process, decoupling it from the training loop. It serves model calls from the harness while recording prompt-response pairs, which are then reconstructed into standard training samples compatible with any RL codebase (e.g., veRL). This is a clever way to bridge the gap between complex, proprietary-like harnesses and generic RL frameworks. 2. **Kubernetes Orchestrator:** Following the design of Orchard, this orchestrator manages the lifecycle of remote containerized rollouts on cloud providers like Microsoft Azure. This enables scalable, elastic execution of rollouts, addressing the challenge that complex harnesses require dedicated, containerized environments that cannot be co-located on training nodes. The paper also details practical considerations for robust operation at scale: * **Asynchronous Rollout and Timeouts:** Imposes wall-clock timeouts on remote rollout jobs to prevent unresponsive rollouts from stalling training, a crucial feature for stability in distributed systems. * **Error Handling:** Discards samples from trajectories that end in non-policy-related errors (e.g., network issues, harness crashes) to avoid injecting misleading training signals. While simple, it's a pragmatic first step. * **Data Synthesis Pipeline:** A significant methodological contribution is the automated pipeline for synthesizing SFT and RL tasks, particularly for data-scarce domains like GUI and computer-use. This pipeline mimics human curation, involving proposal, pruning, environment building, testing with an open LLM/VLM, and refinement. This addresses a major practical challenge in expanding agent research to new domains. Overall, the methodology is sound, practical, and directly tackles the stated problem with a robust system design. It leverages existing technologies (Kubernetes, proxies) in a novel configuration to solve a specific, high-impact ML engineering challenge.
The experimental evaluation is comprehensive, broad, and rigorous, demonstrating the effectiveness and versatility of OpenForgeRL across diverse agentic settings. 1. **Breadth of Environments and Harnesses:** The framework is validated across a wide spectrum: * **Claw Agents (Text-based Tool-use):** Evaluated on ClawEval, QwenClawBench, and MCPAtlas, using various harnesses like ZeroClaw, OpenClaw, and Codex, in addition to a simple loop. * **GUI Agents (Multimodal Browser/Computer-use):** Evaluated on OSWorld-Verified (computer-use), Online-Mind2Web, and WebVoyager (browser-use), using modified Kimi-Agent and Molmo-Web harnesses. This extensive coverage strongly supports the claim of "any harness in any environment." 2. **Strong Empirical Results:** * **Claw Agents:** OpenForgeClaw (30B-A3B MoE) significantly outperforms open baselines of similar size and the untrained backbone model across all three benchmarks (e.g., 31.7 pass^3 on ClawEval, 33.7 on QwenClawBench). The SFT+RL models consistently show substantial improvements over SFT-only, highlighting the efficacy of the end-to-end training. * **GUI Agents:** OpenForgeGUI (8B) achieves superior results on nearly all GUI benchmarks compared to similar-sized models, and impressively matches or surpasses models several times larger (e.g., 63.0 on Online-Mind2Web, 72.3 on WebVoyager, outperforming MolmoWeb trained on 200k tasks with only 2.5k tasks). The consistent gains from RL in this complex multimodal setting are particularly noteworthy. 3. **Valuable Discussion and Analysis:** Beyond benchmark scores, the paper provides insightful analysis enabled by the framework: * **Cross-Harness Comparison:** Reveals that simpler, better-aligned harnesses (e.g., OpenForgeRL's loop, ZeroClaw) are easier to learn and yield higher performance than more complex ones (OpenClaw, Codex), even with advanced features. * **Generalization to Unseen Harnesses:** Demonstrates that training on one harness generalizes to others, and multi-harness training further improves robustness and performance across the board. This is a crucial finding for practical agent development. * **Capabilities Learned by RL:** Detailed behavioral analysis shows that RL improves agentic reliability, such as self-verification, tool coverage, and completing multi-step plans. It also teaches the model to prefer specialized tools over generic ones. This granular insight into *what* RL contributes is highly valuable. 4. **Data Synthesis Validation:** The data synthesis pipeline is shown to be effective in generating useful training data, especially for GUI tasks where data is scarce, enabling the strong performance observed. The experiments are well-designed, the results are compelling, and the analytical discussion adds significant depth, making a strong case for the framework's impact.
The paper makes a strong commitment to reproducibility: * **Code, Data, and Models Release:** The authors explicitly state, "We will release our code, data, and models to facilitate research on harness-based agents." This is the most critical factor for reproducibility. * **Detailed Appendices:** The appendices provide extensive details on training hyperparameters (tab:training-hparams), training curves (fig:claw-curve, fig:computeruse-curve), and a thorough description of the data synthesis pipeline. * **Environment and Harness Details:** Specifics on Kubernetes pod configurations (CPU, RAM), cloud providers (Microsoft Azure), GPU types (B200), and modifications to existing harnesses (e.g., Kimi-Agent, MolmoWeb) are provided. * **Evaluation Protocols:** Clear descriptions of evaluation benchmarks, metrics, and specific protocols (e.g., claim-coverage for MCPAtlas, AgentTrek for Online-Mind2Web) are given. While the code is not yet publicly available, the level of detail provided suggests a strong intent and capability for future reproducibility.
1. **Error Recovery Weakness:** The paper explicitly identifies that "error recovery, however, remains the weakest capability even after RL." This is a significant limitation for agents operating in real-world, noisy environments. The hypothesis that it may require dedicated data or training methods is a good starting point for future work. 2. **Cost of Data Synthesis:** The data synthesis pipeline, while effective, is noted to be "costly in both time and money," especially for RL tasks with robust verifiers (e.g., 16.1 minutes and 4.36 USD per Claw RL task, 21.3 minutes and 6.12 USD per GUI RL task). This could limit its accessibility for researchers without substantial compute budgets. 3. **Engineering-focused Solution:** While a strength in addressing a practical problem, OpenForgeRL is primarily an engineering and systems solution rather than a fundamental algorithmic breakthrough in RL or agent intelligence. Its impact relies on enabling better training of *existing* models and algorithms. 4. **Reliance on External Models for Data Synthesis/Evaluation:** The data synthesis pipeline relies on powerful, often proprietary, LLMs (Claude Opus 4.6, GPT-5.4) for task proposal, pruning, and judging. Similarly, evaluation often uses models like GPT-4o or Gemini 2.5 Pro as judges. This dependency means the quality and cost of the generated data and evaluation are tied to these external services. 5. **Overhead of Proxy/Orchestrator:** While "lightweight," introducing a proxy and Kubernetes orchestrator inherently adds some overhead and complexity compared to a fully integrated, local training setup, though this is a necessary trade-off for the problem being solved.
OpenForgeRL has a substantial broader impact on the field of AI agents: 1. **Democratizing Agent Research:** By providing an open-source framework to train agents in their *real* deployment harnesses, it significantly lowers the barrier for academic researchers and smaller labs to conduct end-to-end training. This directly addresses the "train-deploy mismatch" that has increasingly favored proprietary systems. 2. **Accelerating Agent Development:** Enabling end-to-end training in complex, stateful environments means agents can learn directly from real-world interactions, leading to more robust and capable agents. This could accelerate progress in domains like software engineering, tool use, and multimodal GUI control. 3. **Facilitating Deeper Analysis:** The framework allows researchers to study the impact of harness design and RL training on agent behavior in unprecedented detail, as demonstrated by the paper's discussion section. This can lead to a better understanding of what makes agents effective and how to design better harnesses and training regimes. 4. **New Benchmarking and Data Generation Paradigms:** The automated data synthesis pipeline is a valuable contribution that can help create more diverse and challenging benchmarks for agent evaluation, especially in data-scarce domains. 5. **Bridging the Gap to Frontier Models:** By allowing open models to be trained in the same sophisticated harnesses used by frontier models, OpenForgeRL helps close the capability gap between open and closed-source agent systems. Overall, OpenForgeRL is a timely and impactful contribution that promises to unlock new avenues for research and development in the rapidly evolving field of AI agents. OpenForgeRL introduces a scalable, open-source framework for training harness-based AI agents end-to-end in any environment, bridging the critical gap between complex inference harnesses and standard RL training stacks. This paper presents a robust engineering solution with a lightweight proxy and Kubernetes orchestrator, validated by extensive empirical results across diverse text-based tool-use and multimodal GUI environments, where it outperforms open baselines and provides valuable insights into harness design and the behavioral impact of RL.
An initial high-recall stage in an empirical pipeline decides which items pass to later review, labelling, or modelling, and relevant items it misses are lost to every subsequent stage. We study how many audit labels are needed to certify, with finite-sample validity, that this missed relevant mass is small, and our main results characterise the label complexity of this problem. We first show that no procedure using only labels from inside the candidate set can certify any non-trivial bound on the missed mass: the audit must sample the excluded pool, the only region where unrecovered relevant items can lie. We then prove a matching finite-corpus lower bound. Any valid audit that certifies fewer than $m$ missed relevant items with high probability when none are present, even if adaptive and permitted to label the entire included pool, must inspect on the order of $N_0/m$ excluded-pool labels. Excluded-pool auditing is therefore minimax rate-optimal, not merely convenient, for missed-mass certification in the zero-miss regime. Building on this characterisation, we develop an exact finite-sample toolkit, using binomial and hypergeometric inversion rather than asymptotic approximation, that certifies missed mass, converts it to recall through a two-pool design, certifies pre-specified families of nested candidate generators simultaneously, and produces stress-test certificates against declared perturbation mechanisms. These certificates can be paired with observable review burden to select the least burdensome pre-specified candidate generator meeting a missed-mass target. Every guarantee holds under one discipline: the candidate generator, or the pre-specified family from which it is selected, and the audit rule are fixed before the certification labels are examined.
Primary: London School of Economics and Political Science
All Institutions: London School of Economics and Political Science, small Data Science Institute, small Department of Mathematics, small The Inclusion Initiative
This work has substantial broader impact, particularly in the domain of trustworthy and responsible AI. 1. **Enhanced Trust and Assurance**: Provides a rigorous, finite-sample framework for certifying the performance of critical ML pipeline components, directly contributing to the trustworthiness and reliability of AI systems, especially where false negatives are costly. 2. **Regulatory and Compliance Applications**: The exact, distribution-free guarantees make this framework highly suitable for regulatory audits and compliance in high-stakes domains such as legal discovery, medical diagnostics, financial fraud detection, and systematic reviews, where specific recall targets must be met with high confidence. 3. **Improved ML Pipeline Design**: Offers principled methods for evaluating and selecting candidate generators, leading to more robust and reliable upstream stages in ML pipelines, which in turn can improve the quality of downstream models and analyses. 4. **Resource Optimization**: The bicriteria formulation allows practitioners to select candidate generators that minimize review burden while meeting specified missed-mass targets, leading to more efficient use of human annotation resources. 5. **Foundational Understanding**: The impossibility result and label complexity characterization provide fundamental theoretical insights into the problem of high-recall candidate generation, guiding future research and development in this area. This paper provides a rigorous, finite-sample framework for auditing and certifying high-recall candidate generation systems, establishing fundamental label complexity bounds and offering practical tools for ensuring coverage in critical ML pipelines. The work's strength lies in its theoretical depth, including an impossibility result and minimax optimality for excluded-pool auditing, combined with exact finite-sample statistical guarantees that avoid asymptotic approximations, making it highly relevant for real-world applications where robust certification is paramount.
The paper presents a highly rigorous and theoretically sound methodology for auditing and certifying high-recall candidate generation systems. It operates under two primary models: an i.i.d. population model and a finite-corpus model, allowing for broad applicability. A central methodological contribution is Proposition 1, which formally proves that labels from the included (candidate) pool alone cannot certify any non-trivial bound on missed relevant mass, thereby necessitating auditing the excluded pool. This is a fundamental insight that underpins the entire approach. The paper then provides a matching finite-corpus lower bound (Theorem 1) and establishes the minimax rate-optimality of excluded-pool auditing (Corollary 1) in the zero-miss regime, providing a strong theoretical characterization of the problem's label complexity. The core of the practical toolkit relies on exact finite-sample certification using binomial and hypergeometric inversion, rather than asymptotic approximations. This is a significant strength, as it provides guaranteed coverage even with small audit sample sizes, which is crucial for real-world, high-stakes applications. The methodology extends to various scenarios: 1. **Excluded-pool certification**: Directly bounds missed relevant mass (Theorem 2, 3). 2. **Two-pool recall certification**: Combines excluded-pool and included-pool audits to provide a lower bound on recall (Theorem 4, Corollary 2). 3. **Simultaneous certification**: Allows for auditing pre-specified finite families of candidate generators (e.g., prefix unions) without multiplicity penalties (Theorem 5, Corollary 3). 4. **Operational tools**: Includes sample-size calculations, burden-minimizing stopping rules, and shared-reference designs. 5. **Stress-test certificates**: Provides bounds against declared perturbation mechanisms (Theorem 7). The methodology explicitly enforces a "Design-Certification Separation" (Assumption 1), ensuring the validity of the statistical guarantees by requiring that the candidate generator and audit rule are fixed *before* certification labels are examined. While this imposes a discipline, the paper also discusses valid strategies for iterative design. The assumption of "Noiseless Audit Labels" (Assumption 2) is a simplification, though a limited extension for known sensitivity floors is mentioned. The bicriteria formulation, minimizing review burden subject to a missed-mass constraint, aligns with the Neyman-Pearson approach, providing a practical framework for decision-making.
The provided paper text mentions "A numerical worked example (Section [REF])" but this section is not included in the submission. Consequently, there are no empirical experiments, simulations, or detailed numerical demonstrations presented in the available text. This absence means the practical utility and performance of the proposed certification methods are not empirically validated within the paper. While the theoretical framework is robust, the lack of an experimental evaluation or even a comprehensive numerical example is a significant gap in demonstrating the real-world applicability and behavior of the certificates.
The theoretical results and statistical methods are presented with high mathematical rigor and clarity. The use of standard statistical tools like binomial and hypergeometric inversion means that the core calculations are well-defined and can be implemented by anyone with a solid understanding of statistics. However, without the "numerical worked example" or any accompanying code, a practitioner might find it challenging to directly apply the full toolkit, especially the more complex aspects like simultaneous certification or stress-testing, without further guidance. The mathematical definitions are precise enough for a statistician to reproduce the bounds, but the operational workflow might require more detailed examples.
1. **Lack of Empirical Validation**: The most significant limitation is the absence of empirical experiments or a detailed numerical example in the provided text. This makes it difficult to assess the practical performance, computational overhead, and real-world impact of the proposed methods. 2. **Noisy Audit Labels**: The assumption of noiseless audit labels simplifies the problem. While the paper mentions an extension for known sensitivity floors, a more general treatment of label noise (e.g., unknown noise rates, different noise mechanisms) would be beneficial, as human labeling is inherently prone to error. 3. **Strict Design-Certification Separation**: While essential for validity, this strict separation can be challenging in agile, iterative ML development cycles. Although the paper discusses valid iteration strategies, it still imposes a discipline that might be difficult for some practitioners to adopt without careful planning. 4. **Prevalence Dependence for Recall**: Converting missed-mass certificates to recall certificates often requires a lower bound on the total relevant mass, which may not always be readily available or easy to certify independently, especially in rare-event scenarios. 5. **Computational Cost for Large Families**: While the paper focuses on finite families of candidate generators, the computational cost of simultaneous certification could become a concern for very large or continuously parameterized families, although this is not explicitly discussed.
This work has substantial broader impact, particularly in the domain of trustworthy and responsible AI. 1. **Enhanced Trust and Assurance**: Provides a rigorous, finite-sample framework for certifying the performance of critical ML pipeline components, directly contributing to the trustworthiness and reliability of AI systems, especially where false negatives are costly. 2. **Regulatory and Compliance Applications**: The exact, distribution-free guarantees make this framework highly suitable for regulatory audits and compliance in high-stakes domains such as legal discovery, medical diagnostics, financial fraud detection, and systematic reviews, where specific recall targets must be met with high confidence. 3. **Improved ML Pipeline Design**: Offers principled methods for evaluating and selecting candidate generators, leading to more robust and reliable upstream stages in ML pipelines, which in turn can improve the quality of downstream models and analyses. 4. **Resource Optimization**: The bicriteria formulation allows practitioners to select candidate generators that minimize review burden while meeting specified missed-mass targets, leading to more efficient use of human annotation resources. 5. **Foundational Understanding**: The impossibility result and label complexity characterization provide fundamental theoretical insights into the problem of high-recall candidate generation, guiding future research and development in this area. This paper provides a rigorous, finite-sample framework for auditing and certifying high-recall candidate generation systems, establishing fundamental label complexity bounds and offering practical tools for ensuring coverage in critical ML pipelines. The work's strength lies in its theoretical depth, including an impossibility result and minimax optimality for excluded-pool auditing, combined with exact finite-sample statistical guarantees that avoid asymptotic approximations, making it highly relevant for real-world applications where robust certification is paramount.
As autonomous agents rapidly evolve, their ability to reliably manipulate ubiquitous digital documents has become critical for enabling general-purpose AI assistants and automating complex workspace workflows. In this paper, we introduce DocOps, a deterministically verifiable evaluation framework underpinned by a hierarchical taxonomy that deconstructs document operations inspired by real-world practices into atomic dimensions and escalating workflow complexities. Based on DocOps, we systematically evaluate representative closed- and open-source models across various agentic harnesses, revealing that even the most advanced frontier configurations still exhibit profound limitations when handling highly coupled, long-range tasks. Furthermore, a fine-grained analysis of existing agents' manipulation behaviors uncovers 3 key failure modes: long-term state tracking collapse, shallow semantic verification, and destructive editing of structural metadata. Ultimately, our work exposes the capability boundaries of agents in maintaining global document consistency, shedding light on the future design of robust, non-destructive agents for complex digital ecosystems.
Primary: Not specified in the provided text (placeholder 'Address line')
All Institutions: Not specified in the provided text (placeholder 'Address line')
DocOps makes a significant contribution to the field of autonomous agents and general-purpose AI. By providing a rigorous, verifiable benchmark for complex document operations, it addresses a critical gap in evaluating agents' ability to interact with ubiquitous digital documents. The findings expose fundamental limitations of current frontier models in maintaining global document consistency and avoiding destructive modifications, shifting research focus from isolated tool invocation to state-aware, non-destructive agent design. The identified failure modes offer clear diagnostic targets for improving agent architectures, planning mechanisms, and verification capabilities. This work will likely guide the development of more robust AI assistants for workspace automation, impacting productivity across various industries. The benchmark itself is poised to become a standard tool for researchers and practitioners, fostering innovation in a crucial area of human-computer interaction. DocOps introduces a rigorously verifiable evaluation framework and benchmark for autonomous agents performing complex, stateful document operations, revealing significant limitations of current frontier models in maintaining global consistency and avoiding destructive edits. This paper makes a substantial technical contribution by defining a novel taxonomy for document manipulation, developing a deterministic artifact-level verification system, and conducting a comprehensive empirical evaluation that uncovers critical failure modes and provides actionable insights for the design of future robust, non-destructive agents.
The methodology for DocOps is exceptionally well-conceived and rigorously designed. The core contribution is a deterministically verifiable evaluation framework for autonomous agents performing complex document operations. A key strength is the hierarchical taxonomy, which deconstructs document operations along two orthogonal axes: atomic capabilities (content, format, structure) and workflow depth (L1-L4). This allows for fine-grained diagnosis of agent failures, moving beyond coarse task-level success metrics. The task construction pipeline is robust, involving seed collection, formalization, source-artifact synthesis, and iterative human review, ensuring practical relevance, clarity, and consistency across 210 tasks. Crucially, DocOps introduces a novel deterministic verifier that directly inspects output files using native document libraries. This verifier employs three types of predicates (structural, linguistic, preservation) to not only check task completion but also to ensure structural validity and preservation of out-of-scope elements, addressing a major limitation of prior benchmarks. The fidelity of this verifier is rigorously assessed through a human audit and mutation-based stress test, demonstrating high agreement. The evaluation protocol, utilizing the Harbor framework, is standard and well-defined, ensuring reproducibility. Overall, the methodology is a significant advancement in benchmarking agent capabilities for complex, stateful digital environments.
The experimental evaluation is comprehensive and insightful. The paper systematically evaluates a diverse set of models, including leading closed-source (GPT-5.5, GPT-5.4, Claude Sonnet 4.6) and open-source (DeepSeek-V4-Pro, Qwen, Gemma, GLM) LLMs. These models are tested across four distinct agentic harnesses (DocTools, Terminus-2, Claude Code, Codex) representing different interface regimes, and with/without explicit skill injection. This broad coverage provides a holistic view of current agent capabilities. The results reveal profound limitations: even the most advanced frontier configuration (GPT-5.5 with Codex and skills) achieves only a 0.671 pass rate, which drops sharply on workflow-level (L3/L4) tasks. This is a significant empirical finding. The detailed analysis further uncovers that workflow difficulty is highly dependent on the *coupling* of underlying document states (e.g., Excel tasks degrade much more severely than PDF tasks), rather than just the number of operations. The paper also identifies and quantifies three pervasive failure modes: long-term state tracking collapse, shallow semantic verification, and destructive editing of structural metadata, with semantic verification gaps being the most dominant. The analysis of harness impact shows that open-ended programming environments generally outperform constrained tool use, and that skills offer non-uniform benefits, sometimes even increasing cost without significant performance gains for frontier models. The experiments are well-designed, the results are clearly presented (tables, figures), and the findings provide actionable insights for future agent development.
Reproducibility is a strong suit of this paper. The authors explicitly state that "Both the code and dataset are publicly available: https://github.com/icip-cas/DocOps". Each task is packaged as a self-contained Harbor bundle, including source artifacts, natural language instructions, optional skills, and the deterministic verifier. This packaging, combined with the use of the Harbor framework for execution, greatly facilitates replication of the experiments. The detailed descriptions of the task construction pipeline, verifier design, and experimental protocol (including model IDs and access routes in the appendix) further enhance reproducibility. The deterministic nature of the verifier, independent of LLM-as-a-judge, is a critical factor in ensuring consistent evaluation outcomes.
The authors acknowledge several limitations. DocOps focuses on deterministic, offline document-editing tasks, thus not covering workflows requiring live external services, collaborative editing, or interactive user clarification. The benchmark currently contains 210 tasks, and scaling it is labor-intensive due to the need for structurally valid artifacts, clear editing scopes, and manual review. This limits the current breadth of document domains and workflow complexities. Finally, token-cost comparisons across harnesses should be interpreted with care due to varying fidelity in usage statistics exposed by different agent runtimes. These are reasonable limitations for a novel and complex benchmark, and the authors outline plans for future expansion.
DocOps makes a significant contribution to the field of autonomous agents and general-purpose AI. By providing a rigorous, verifiable benchmark for complex document operations, it addresses a critical gap in evaluating agents' ability to interact with ubiquitous digital documents. The findings expose fundamental limitations of current frontier models in maintaining global document consistency and avoiding destructive modifications, shifting research focus from isolated tool invocation to state-aware, non-destructive agent design. The identified failure modes offer clear diagnostic targets for improving agent architectures, planning mechanisms, and verification capabilities. This work will likely guide the development of more robust AI assistants for workspace automation, impacting productivity across various industries. The benchmark itself is poised to become a standard tool for researchers and practitioners, fostering innovation in a crucial area of human-computer interaction. DocOps introduces a rigorously verifiable evaluation framework and benchmark for autonomous agents performing complex, stateful document operations, revealing significant limitations of current frontier models in maintaining global consistency and avoiding destructive edits. This paper makes a substantial technical contribution by defining a novel taxonomy for document manipulation, developing a deterministic artifact-level verification system, and conducting a comprehensive empirical evaluation that uncovers critical failure modes and provides actionable insights for the design of future robust, non-destructive agents.
Bayesian online learning promises uncertainty-aware prediction on data streams, but its performance hinges on inferential choices, including learning rates, prior distributions and variational families, which are usually fixed before seeing the stream. We address this by treating Bayesian update rules as experts and aggregating the Bayesian experts according to sequential predictive losses. We prove that the resulting aggregate competes with the best expert in hindsight at an aggregation cost determined by how each expert's per-round performance is evaluated. We instantiate the framework in online conformal inference and Gaussian process regression. The conformal inference application yields a smoothed Bayesian counterpart of adaptive conformal inference with long-run randomized coverage, while the Gaussian process application gives an oracle inequality in cumulative predictive Kullback-Leibler risk and adaptation to unknown Hölder smoothness up to logarithmic factors. Experiments show that the aggregate tracks strong experts without oracle expert selection.
Primary: Inha University
All Institutions: Inha University
This work has significant broader impact for the field of online learning and Bayesian methods. It provides a principled and modular framework for making Bayesian online learning more adaptive and robust to critical inferential choices (learning rates, priors, variational families) that are often fixed arbitrarily. This can lead to: * **More reliable uncertainty quantification:** By adapting inferential choices, the resulting predictions and uncertainty estimates are less sensitive to misspecification or nonstationarity. * **Reduced hyperparameter tuning:** The aggregation mechanism automates the selection of optimal inferential settings, reducing the need for manual tuning. * **Enhanced adaptivity:** The framework allows Bayesian models to adapt to changing data stream characteristics (e.g., smoothness, noise levels, nonstationarity) in a theoretically grounded manner. * **New theoretical insights:** The distinction between mean-loss and annealed-loss aggregation offers fundamental insights into how to evaluate and combine Bayesian predictions, with implications for other areas of online learning. The applications to conformal inference and Gaussian processes demonstrate its utility in practical, high-impact areas where robust uncertainty quantification and adaptation are crucial. This paper introduces a novel expert-aggregation framework for adaptive Bayesian online learning, distinguishing between mean-loss and annealed-loss aggregation with corresponding $O(T)$ and $O(\log K)$ regret bounds, and demonstrates its effectiveness in online conformal inference and Gaussian process regression with adaptation to unknown Hölder smoothness. The work provides a principled and modular approach to address the sensitivity of Bayesian online learning to fixed inferential choices, offering strong theoretical guarantees and comprehensive empirical validation across diverse online learning settings, thereby advancing the robustness and adaptivity of uncertainty-aware prediction systems.
The paper proposes a novel two-level framework for adaptive Bayesian online learning, treating different Bayesian update rules (experts) as distribution-valued entities and aggregating their posterior predictive distributions. The core methodological contribution lies in identifying two distinct ways to evaluate these Bayesian experts: mean-loss aggregation and annealed-loss aggregation. The authors rigorously derive regret bounds for both, showing that mean-loss aggregation generally incurs an $O(T)$ cost, while annealed-loss aggregation achieves a much faster $O(\log K)$ cost. This distinction is crucial, as it links the choice of evaluation metric to the statistical properties of the target functional (e.g., posterior mean vs. full predictive distribution). The framework is modular, allowing various Bayesian online learning algorithms (SVB, OGA) to serve as base experts. The paper then instantiates this general framework in two significant applications: online conformal inference (yielding Bayes-ACI and Bayes-DtACI) and online Gaussian process regression with unknown smoothness. For the latter, the annealed-loss aggregation is shown to adapt to unknown Hölder smoothness at minimax rates up to logarithmic factors, effectively replacing a hierarchical prior with sequential aggregation. The theoretical development is sound, leveraging established results from prediction with expert advice and extending them to the Bayesian online learning context. The discussion on the curvature of loss functions and its impact on regret rates is particularly insightful.
The experimental evaluation is comprehensive and well-designed, covering three distinct online learning scenarios. 1. **Online Variational Benchmarks:** The paper evaluates the proposed expert aggregation methods (SVB-EA, OGA-EA, OGD-EA) on standard binary classification and regression datasets. Results demonstrate that the adaptive aggregates consistently track the performance of the best fixed expert in hindsight, which is a strong indicator of successful adaptation. The sensitivity of individual experts to learning rates is clearly shown, highlighting the value of aggregation. 2. **Online Conformal Inference:** Bayes-DtACI is tested in a nonstationary setting with abrupt changes in residual scale and heavy-tailed errors. It is compared against the original DtACI. Bayes-DtACI shows more stable cumulative coverage, especially under Student-$t$ errors, suggesting improved robustness due to the Gaussian-smoothed updates. The rolling coverage also adapts smoothly to regime changes. 3. **Online GP Regression:** The annealed-loss aggregation for GPs is evaluated across three stationary settings with varying function smoothness/length scales, and a nonstationary setting. The aggregate consistently tracks the best fixed bandwidth expert in terms of cumulative negative log-likelihood. Crucially, the aggregation weights dynamically reallocate, adapting to the effective smoothness of the underlying function, which is a key theoretical claim. In the nonstationary setting, GP-EA remains competitive with strong online regression baselines and demonstrates effective adaptation to regime changes by reallocating weights. Overall, the experiments provide strong empirical evidence for the theoretical claims, demonstrating both the adaptivity and robustness of the proposed framework across diverse applications. The choice of metrics and baselines is appropriate.
The paper provides a good level of detail for reproducibility. The algorithms (Alg. 1) are clearly described. Expert specifications, meta-learning rates, and sharing parameters are detailed for each experiment. The use of specific software (Python package River, version 0.20.1) is mentioned. Experiments are repeated 30 times. While specific code is not provided (common for arXiv preprints), the methodological and experimental descriptions are sufficiently thorough for a skilled researcher to reproduce the main results.
The authors acknowledge several limitations in their future work section. 1. **Fixed-share aggregation:** The current framework primarily uses fixed-share aggregation. More advanced, strongly adaptive, or parameter-free aggregation methods could offer sharper guarantees and faster adaptation to nonstationary streams. 2. **Finite, prespecified expert collection:** The theory is developed for a finite, pre-specified grid of experts. Extending this to continuous, data-dependent, or growing expert families would be more powerful but would require controlling statistical complexity and computational cost. 3. **Computational Cost:** For a large number of experts $K$, maintaining $K$ separate Bayesian updates and their aggregation can be computationally intensive, especially for complex models like GPs. 4. **Known Noise Level in GP Theory:** The theoretical analysis for GP regression assumes a known noise level, which is not practical. While the experiments address this by aggregating over a product grid of bandwidths and noise levels, the theoretical guarantees for this extended setting are not explicitly derived, though the authors suggest it's covered by lifting the parameter space.
This work has significant broader impact for the field of online learning and Bayesian methods. It provides a principled and modular framework for making Bayesian online learning more adaptive and robust to critical inferential choices (learning rates, priors, variational families) that are often fixed arbitrarily. This can lead to: * **More reliable uncertainty quantification:** By adapting inferential choices, the resulting predictions and uncertainty estimates are less sensitive to misspecification or nonstationarity. * **Reduced hyperparameter tuning:** The aggregation mechanism automates the selection of optimal inferential settings, reducing the need for manual tuning. * **Enhanced adaptivity:** The framework allows Bayesian models to adapt to changing data stream characteristics (e.g., smoothness, noise levels, nonstationarity) in a theoretically grounded manner. * **New theoretical insights:** The distinction between mean-loss and annealed-loss aggregation offers fundamental insights into how to evaluate and combine Bayesian predictions, with implications for other areas of online learning. The applications to conformal inference and Gaussian processes demonstrate its utility in practical, high-impact areas where robust uncertainty quantification and adaptation are crucial. This paper introduces a novel expert-aggregation framework for adaptive Bayesian online learning, distinguishing between mean-loss and annealed-loss aggregation with corresponding $O(T)$ and $O(\log K)$ regret bounds, and demonstrates its effectiveness in online conformal inference and Gaussian process regression with adaptation to unknown Hölder smoothness. The work provides a principled and modular approach to address the sensitivity of Bayesian online learning to fixed inferential choices, offering strong theoretical guarantees and comprehensive empirical validation across diverse online learning settings, thereby advancing the robustness and adaptivity of uncertainty-aware prediction systems.
Natural-language autoencoders score explanations of hidden activations by reconstruction: an explanation is deemed faithful if the activation can be regenerated from it. The test is structurally insensitive to individual false claims: if flipping a claim does not change the reconstruction, the claim is never penalized. We show the test is passed in two ways, neither faithful. On a released Qwen-2.5-7B verbalizer, explanations reconstruct well above chance while ~2% of specific claims are reconstruction-dependent, so the score tracks gist, not specific facts. Under exact synthetic ground truth, the standard recipe develops co-adapted private codes (false wording the reconstruction depends on) in 5/5 runs, and fixes that leave the target model unchanged do not help. We contribute two audit protocols, the grounded-vs-true cross and the evaluator swap, and RECAP (Readable Encodings via Co-trained Auxiliary Predictors): linear heads trained alongside the target model to keep designated content decodable. On RECAP-trained sandbox models, fresh verbalizers state the designated content truly and the codes vanish, at a +0.001-nat cost. This replicates on a pretrained Pythia-160M: the content becomes reliably probe-decodable, though a fresh verbalizer conveys it only in part (truth 0.44-0.46 vs a near-zero control). For interpretability, high reconstruction does not certify individual claims. For AI safety, RECAP makes designated internal content independently checkable against probes rather than asserted by prose a model can game: an independent probe scores the verbalizer's true claims above its false ones (AUC 0.96, vs 0.82 without RECAP). Against an adversary that edits an explanation to maximize the reconstruction score while lying (suppressing ~87% of its lie penalty), the RECAP probe still flags the lies (AUC 0.95) while the control probe collapses to chance (0.51).
Primary: Unknown
All Institutions: Unknown
This paper provides a critical audit of activation explanation methods and proposes RECAP, a training-time intervention that ensures internal content is probe-decodable, thereby enabling reliable, independent verification of model claims against adversarial gaming.
The paper introduces a rigorous critique of Natural-Language Autoencoders (NLAs) used for activation explanation. It identifies two critical failure modes: "gist" reliance (where explanations reconstruct well but contain ungrounded specific claims) and "private codes" (where the reconstructor relies on specific, non-factual wording co-adapted during training). To address the latter, it proposes RECAP (Readable Encodings via Co-trained Auxiliary Predictors), a method that trains auxiliary linear heads on the target model to ensure designated content remains probe-decodable. This shifts the burden of faithfulness from the verbalizer's reconstruction score to the model's internal representational geometry. The methodology is sound, leveraging synthetic sandboxes for exact ground truth and real models for scalability checks.
The experimental design is comprehensive and multi-layered. It begins with an audit of a released Qwen-2.5-7B NLA, demonstrating high reconstruction scores despite low claim-level grounding. It then moves to a synthetic sandbox with exact ground truth to isolate "private codes," showing they emerge consistently in standard NLA training. The core contribution, RECAP, is validated in the sandbox (showing 100% decodability and truth) and scaled to Pythia-160M (showing significant improvements in probe-decodability and verbalizer truthfulness compared to controls). Crucially, the paper includes an adversarial test where an agent attempts to game the reconstruction score; RECAP's probe-based monitor successfully flags lies while the control probe fails. The use of independent fresh probes to evaluate decodability prevents circularity.
The paper provides detailed protocols for the synthetic domains, including grammar specifications and slot definitions. It mentions the use of public models (Qwen-2.5-7B, Pythia-160M) and standard training setups. The technical appendix (referenced but not fully included in the text provided) likely contains hyperparameters and seeds. The clear distinction between co-trained heads and independent evaluation probes ensures that results are not artifacts of the reader. The synthetic nature of the primary validation experiments makes them highly reproducible.
The primary limitation is the scale of the model experiments; RECAP is demonstrated on Pythia-160M, which is small compared to frontier models. The paper acknowledges that decodability does not guarantee verbalizability; the verbalizer still struggles to convey all decoded content faithfully. Additionally, RECAP requires co-training, meaning it cannot be easily applied to frozen, pre-trained models without significant retraining costs. The "target design" rule for auxiliary heads is heuristic and may be difficult to optimize for complex, open-ended content.
This work has significant implications for AI safety and interpretability. By demonstrating that reconstruction-based faithfulness tests are structurally insufficient, it warns practitioners against trusting NLA scores as proxies for truth. RECAP offers a pathway to making internal model states independently verifiable, which is crucial for oversight of capable agents. It shifts the paradigm from "explaining the model" to "making the model explainable" by design. This paper provides a critical audit of activation explanation methods and proposes RECAP, a training-time intervention that ensures internal content is probe-decodable, thereby enabling reliable, independent verification of model claims against adversarial gaming.
Scaling executable agent training data for LLM post-training is bottlenecked by substrate-bound methods that tie task generation to predefined tools, repositories, or skill graphs: expanding coverage requires manual substrate engineering, each new domain demands a bespoke pipeline, and the resulting task distributions often reflect substrate biases rather than real-world demand. We introduce NexForge, a requirement-driven framework that takes high-level capability requirements as input and synthesizes diverse, executable agent tasks and expert trajectories for SFT. NexForge first investigates real-world demand to construct representative scenarios and task profiles, then performs distribution-aware compilation to generate task directives. For each directive, NexForge automatically retrieves or constructs the required files, dependencies, and runtime configurations, and finally synthesizes expert rollouts and produces training trajectories. Without domain-specific infrastructure, NexForge produces 3.6K terminal and 2K office tasks, improving Qwen3.5-35B-A3B Base from 22.5\% to 52.0\% on Terminal-Bench 2.0 and from 813 to 1338 Elo on GDPval; scaling further to 43.2K terminal tasks yields 58.4\%, on par with Claude Opus 4.6 equipped with Claude Code. Scaled further, NexForge-synthesized data contributes to the training of Nex-N2, a family of publicly available agent models that lift Qwen3.5-35B-A3B to 75.3\% on Terminal-Bench 2.1 and to 1585 Elo on GDPval -- achieving state-of-the-art open-source performance and surpassing several frontier proprietary systems. Nex-N2 models are available at https://nex.sii.edu.cn/.
Primary: SII (Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences)
All Institutions: SII (Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences)
NexForge presents a compelling and effective pipeline for scaling agent training data through requirement-driven synthesis, demonstrating that high-quality, diverse task generation can significantly boost LLM agent performance, achieving state-of-the-art open-source results on key benchmarks.
The paper introduces NexForge, a framework designed to address the data bottleneck in training LLM-based agents. The core innovation lies in shifting from "substrate-bound" task generation (which relies on predefined tools or codebases) to a "requirement-driven" approach. The methodology involves three key stages: 1) Analyzing real-world demand to create representative scenarios and task profiles; 2) Distribution-aware compilation to generate high-level task directives; and 3) Automatic synthesis of executable environments (files, dependencies, runtime configs) and expert rollouts for Supervised Fine-Tuning (SFT). This approach aims to reduce manual engineering and mitigate substrate biases. The method is technically sound, leveraging existing LLM capabilities for code generation and environment setup, but the novelty is incremental rather than foundational. It represents a sophisticated engineering pipeline rather than a new algorithmic breakthrough.
The experimental section demonstrates significant empirical improvements. Using Qwen3.5-35B-A3B as the base, the authors show a jump from 22.5% to 52.0% on Terminal-Bench 2.0 and from 813 to 1338 Elo on GDPval with 3.6K terminal and 2K office tasks. Scaling to 43.2K terminal tasks pushes performance to 58.4%, which is comparable to Claude Opus 4.6 with Claude Code. Furthermore, the synthesized data is used to train "Nex-N2," achieving state-of-the-art open-source results (75.3% on Terminal-Bench 2.1, 1585 Elo on GDPval). The results are impressive and suggest that high-quality, diverse, requirement-driven data is a critical lever for agent performance. The evaluation is rigorous, covering multiple benchmarks and comparing against strong proprietary baselines.
The paper provides a project URL (https://nex.sii.edu.cn/) which likely contains code and model weights. The description of the pipeline (requirement analysis -> directive compilation -> environment synthesis -> rollout) is detailed enough to be reproducible by a team with sufficient resources. However, the "expert rollouts" likely rely on a strong teacher model or human-in-the-loop, which can introduce variability. The specific "distribution-aware compilation" algorithm is not fully detailed in the abstract, so full reproducibility depends on the completeness of the main text and code release.
The paper does not explicitly discuss the cost of generating 43.2K high-quality tasks, which can be significant. The reliance on a "requirement-driven" approach assumes that high-level requirements can be effectively mapped to executable tasks, which may fail in domains with ambiguous or complex implicit constraints. Additionally, the "substrate biases" argument, while valid, might be overstated if the underlying LLMs themselves have biases in their training data that NexForge cannot correct. The evaluation is primarily on coding/terminal tasks; generalization to other agent domains (e.g., web browsing, multi-modal reasoning) is not demonstrated.
This work has significant implications for the democratization of capable AI agents. By providing a scalable method for generating high-quality training data, it lowers the barrier to entry for developing specialized agents. The release of Nex-N2 models contributes to the open-source ecosystem. However, the potential for misuse (e.g., generating malicious code or automating cyberattacks) is a concern that should be addressed in the broader impact statement. The success of such frameworks may accelerate the arms race in agent capabilities, raising safety and alignment challenges. NexForge presents a compelling and effective pipeline for scaling agent training data through requirement-driven synthesis, demonstrating that high-quality, diverse task generation can significantly boost LLM agent performance, achieving state-of-the-art open-source results on key benchmarks.
Large-scale visual generators are increasingly capable but costly to train, fine-tune, and deploy. We introduce Mage-Flow, a compact 4B-scale generative stack for efficient text-to-image generation and instruction-based image editing. The stack is built from two co-designed components: Mage-VAE, a lightweight high-fidelity latent tokenizer, and a Native-Resolution Multimodal Diffusion Transformer trained with rectified flow matching. Mage-VAE uses one-step diffusion-style encoding and decoding with anchor-latent regularization, preserving the reconstruction quality of strong public VAEs while reducing tokenization cost by more than an order of magnitude. Together with native-resolution packing and stack-level CUDA kernel fusion, the stack supports flexible-resolution training and improves end-to-end training throughput by about 2.5times. Built on this foundation, we develop a complete model family with Base, RL-aligned, and Turbo variants for both generation and editing. Diffusion-NFT improves prompt following, text rendering, aesthetic quality, and editing fidelity, while few-step distillation with adversarial perceptual guidance produces 4-step Turbo models for low-latency inference. Despite its compact scale, Mage-Flow and Mage-Flow-Edit achieves competitive performance across standard generation and editing benchmarks. More importantly, the Turbo variants make high-resolution generation and editing practical for interactive use: at 1024^2 resolution on a single NVIDIA A100 GPU, Mage-Flow-Turbo generates an image in 0.59s, and Mage-Flow-Edit-Turbo edits an image in 1.02s, while maintaining a small memory footprint. These results show that careful tokenizer--backbone--system co-design can deliver strong high-resolution generation and editing within an efficient 4B model family.
Primary: Microsoft
All Institutions: Microsoft
Mage-Flow presents a significant engineering and methodological contribution by co-designing a lightweight VAE and native-resolution diffusion transformer, achieving state-of-the-art efficiency for high-resolution image generation and editing on consumer-grade hardware, thereby democratizing access to powerful generative AI tools.
The paper proposes a co-designed generative stack, Mage-Flow, consisting of a lightweight VAE (Mage-VAE) and a native-resolution diffusion transformer. The novelty lies in the system-level co-design: using one-step diffusion-style encoding/decoding with anchor-latent regularization to drastically reduce tokenization overhead, combined with native-resolution packing and CUDA kernel fusion to enable efficient training and inference. This approach addresses the computational bottlenecks of high-resolution image generation. The use of rectified flow matching and the development of specific variants (Base, RL-aligned, Turbo) for different use cases (generation vs. editing, speed vs. quality) demonstrates a comprehensive engineering and methodological effort.
The authors present a model family including Base, RL-aligned, and Turbo variants. They report competitive performance on standard generation and editing benchmarks. Crucially, they highlight inference efficiency: 0.59s for generation and 1.02s for editing at 1024^2 resolution on a single A100 GPU. These metrics are significant for practical deployment. The evaluation covers both quality (prompt following, text rendering, aesthetics) and efficiency (latency, memory footprint), providing a robust assessment of the trade-offs.
The paper provides code, models, and a project page, which strongly supports reproducibility. The description of the architecture (Mage-VAE, Native-Resolution DiT) and training techniques (rectified flow, adversarial perceptual guidance) is detailed enough for other researchers to attempt replication, assuming access to similar computational resources.
As a 4B parameter model, it may still lag behind larger foundation models (e.g., Flux, SD3, DALL-E 3) in terms of absolute peak quality or complex semantic understanding, although the paper claims competitiveness. The "native-resolution" approach, while efficient, may still face challenges with extremely high resolutions or complex multi-subject compositions compared to models specifically optimized for those edge cases. The reliance on specific CUDA kernel fusion optimizations might limit portability to non-NVIDIA hardware.
By making high-resolution, interactive image generation and editing accessible on single GPUs, this work lowers the barrier to entry for developers and researchers. It promotes more sustainable AI by reducing the energy and hardware costs associated with training and inference. The focus on editing also has implications for creative workflows and content creation industries. Mage-Flow presents a significant engineering and methodological contribution by co-designing a lightweight VAE and native-resolution diffusion transformer, achieving state-of-the-art efficiency for high-resolution image generation and editing on consumer-grade hardware, thereby democratizing access to powerful generative AI tools.
LLM agent failures are difficult to debug because the step where an error surfaces is often not the one that caused it. Existing observability tools replay execution traces but provide little support for identifying the root cause or translating diagnosis into recovery. We present AgentDebugX, an open-source debugging framework that organizes debugging as a closed loop of Detect, Attribute, Recover, and Rerun. At its core, DeepDebug performs multi-turn root-cause diagnosis through global trajectory understanding, structure-guided investigation, and cross-examination. On the Who and When benchmark, DeepDebug achieves the best strict attribution accuracy among the evaluated methods on both tested open-weight backbones, reaching 28.8 percent exact agent-and-step accuracy on qwen3.5-9b versus 21.7 percent for the strongest single-pass baseline. On GAIA, DeepDebug repairs 13 of 73 failed tasks in a single rerun, compared with 4 to 6 for three decoupled self-correction baselines, improving overall accuracy from 55.8 percent to 63.6 percent. AgentDebugX exposes this workflow through a Python library, CLI, web console, and installable agentic skill, and provides an opt-in Error Hub for sharing scrubbed failure-diagnosis-repair bundles and reusing them as debugging memory.
Primary: Stanford University
All Institutions: Stanford University
AgentDebugX presents a significant step forward in LLM agent observability by introducing a structured, closed-loop debugging framework that connects root-cause attribution with actionable recovery, demonstrating measurable improvements in agent repair rates on complex benchmarks.
The paper proposes AgentDebugX, a comprehensive framework for LLM agent debugging that formalizes the process as a closed loop of Detect, Attribute, Recover, and Rerun. The core methodological contribution is "DeepDebug," a multi-turn root-cause diagnostic agent. Unlike single-pass attribution methods that often fail on long-horizon traces, DeepDebug employs a structured investigation strategy: a global read followed by a structure-guided probe (bisecting for single agents, tracing handoffs for multi-agent systems) and a cross-examination phase to arbitrate between conflicting hypotheses. This approach addresses the specific challenge of "latent" errors where the symptom surface is far from the root cause. The framework also introduces a portable trajectory representation and an "Error Hub" for sharing scrubbed failure-diagnosis-repair bundles, aiming to create a cumulative debugging memory. The methodology is sound and addresses a genuine gap in the current agent observability landscape, which largely focuses on logging rather than actionable diagnosis and recovery.
The evaluation is split into two parts: attribution accuracy on the Who&When benchmark and end-to-end recovery on GAIA. On Who&When, DeepDebug achieves 28.8% strict agent-and-step accuracy on Qwen3.5-9b, outperforming the strongest single-pass baseline (21.7%). While this is a relative improvement, the absolute accuracy remains low, highlighting the difficulty of the task. On GAIA, the framework repairs 13 of 73 failed tasks in a single rerun, compared to 4-6 for decoupled self-correction baselines, improving overall accuracy from 55.8% to 63.6%. The experiments are well-controlled, comparing against relevant baselines (Reflexion, CRITIC, AutoManual) and providing ablations on the diagnostic turns. However, the GAIA evaluation is limited to a single policy model and a single rerun, which may overestimate the generalizability of the recovery gains. The attribution gains are modest in absolute terms, suggesting that while the method is state-of-the-art among evaluated approaches, the problem of automated root-cause analysis for LLM agents remains unsolved.
The paper provides an open-source toolkit, a Python library, and detailed prompts for the diagnostic agents. The code is available on GitHub, and the paper includes specific details on the trace schema and evaluation protocols. The use of standard benchmarks (Who&When, GAIA) enhances reproducibility. The inclusion of an "Error Hub" format specification also aids in future reproducibility and comparison.
The authors acknowledge several limitations. The evaluation does not measure developer debugging time or UI usability, which are critical for practical adoption. The attribution gains are model-dependent, with the multi-turn approach showing less benefit on stronger hosted models (GPT-5.4-mini, Gemini-3.5-flash) where single-pass reading is already effective. The GAIA experiment evaluates the full recipe rather than isolating the effect of attribution alone. The Error Hub and taxonomy induction features are implemented but not yet evaluated. The scrubber for sensitive data is pattern-based and may not catch all PII.
AgentDebugX has the potential to make agent reliability more inspectable and measurable, moving beyond proprietary black-box debugging. By providing an open-source toolkit and a shared format for failure cases, it lowers the barrier for researchers and smaller organizations to study robustness. The Error Hub concept could foster a community-driven corpus of agent failures, accelerating progress in agent reliability. However, the collection and sharing of agent traces raise privacy and security concerns, which the paper addresses through opt-in sharing and redaction mechanisms. AgentDebugX presents a significant step forward in LLM agent observability by introducing a structured, closed-loop debugging framework that connects root-cause attribution with actionable recovery, demonstrating measurable improvements in agent repair rates on complex benchmarks.
Summation error depends on partial-sum order, which standard worst-case bounds omit. To capture this dependence, we derive an exact mean-square error (MSE) recurrence for a binary reduction tree T under conditionally unbiased rounding. With unit roundoff u, the constant-nu model sets the local variance at pre-rounding value x to nu u^2 x^2. Its leading tree-dependent cost for the input vector p is p^T K_T p, where the common-ancestor kernel K_T counts the internal ancestors shared by each pair of leaves. For i.i.d. inputs of mean mu and variance tau^2, this expected cost is tau^2 Lambda_1(T) + mu^2 Lambda_2(T), where Lambda_1 is total leaf depth and Lambda_2 sums squared internal-subtree sizes; Lambda_1 governs centered inputs, while Lambda_2 captures nonzero means. We use these statistics to characterize optimal tree topologies and schedules. Balanced and sequential trees attain the centered extrema. For k inputs, optimal two-stage sequential blocking yields root-mean-square (RMS) error scaling as k^{3/4}. For fixed-stage hierarchies, geometric schedules are optimal for centered inputs, whereas the optimal noncentered stage exponents halve successively. For independent centered inputs with unequal variances, Huffman coding minimizes variance-weighted depth over free leaf assignments. We extend the kernel to matrix multiplication through operand Gram matrices. We then test the approximation under round-to-nearest using exact residuals. Across binary64, binary32, and software-emulated binary16 and bfloat16, the model recovers the ordering among tree topologies; K_T tracks AR(1) partial-sum costs. For GEMM, independently calibrated predictions differ from measurements by at most 3% on the tested grid. A reduction tree extracted from an array library predicts the measured RMS scaling. However, stagnation and bias in positive low-precision sums limit the model's applicability.
Primary: Oak Ridge National Laboratory
All Institutions: Oak Ridge National Laboratory
This paper provides a rigorous second-moment theory for floating-point reduction trees, deriving exact error recurrences and optimal topologies that significantly advance the understanding of numerical stability in parallel reductions, with direct applications to improving the accuracy of HPC and ML libraries.
The paper presents a rigorous theoretical framework for analyzing floating-point reduction errors, specifically focusing on the variance of partial sums in reduction trees. The core methodological contribution is the derivation of an exact mean-square error (MSE) recurrence relation for binary reduction trees under the conditionally unbiased rounding model. The authors introduce the "common-ancestor kernel" $K_T$, which quantifies the structural impact of tree topology on error propagation. They decompose the expected cost into terms dependent on input variance ($\Lambda_1$) and mean ($\Lambda_2$), allowing for the characterization of optimal tree topologies (e.g., balanced vs. sequential, Huffman coding for unequal variances). The methodology extends naturally to matrix multiplication via Gram matrices, providing a unified view of reduction error in linear algebra operations. The approach is mathematically sound, leveraging statistical properties of floating-point arithmetic rather than worst-case bounds, which offers a more realistic model for typical workloads.
The authors validate their theoretical model through extensive experiments across multiple precision formats (binary64, binary32, software-emulated binary16, and bfloat16). They demonstrate that the model accurately predicts the ordering of RMS error among different tree topologies and tracks the costs of autoregressive (AR(1)) processes. For General Matrix Multiplication (GEMM), the model's predictions differ from measurements by at most 3% on the tested grid. They also test a reduction tree extracted from an actual array library, confirming the model's predictive power for real-world implementations. The experiments are well-designed, covering both synthetic i.i.d. inputs and structured data, and effectively bridge the gap between theoretical bounds and empirical behavior.
The paper provides detailed mathematical derivations and specifies the rounding models and input distributions used in experiments. The mention of "software-emulated" formats suggests that the authors have implemented or utilized existing tools for lower-precision simulation, which aids reproducibility. However, the paper does not explicitly provide a link to the source code or specific software versions used for the simulations in the abstract or main text provided. Given the institutional context (ORNL) and the nature of the work, code is likely available or reproducible, but explicit URLs are missing from the provided text.
The authors explicitly acknowledge limitations, noting that the model's applicability is limited by stagnation and bias in positive low-precision sums. This suggests that the conditionally unbiased rounding assumption may break down in specific edge cases, particularly with low-precision formats like bfloat16 or binary16 where dynamic range and precision constraints are tighter. The model is primarily statistical (expectation/variance) and may not capture worst-case scenarios or specific pathological inputs that trigger catastrophic cancellation or overflow in ways not captured by the second-moment analysis.
This work has significant implications for the design of high-performance computing (HPC) libraries and machine learning frameworks that rely heavily on parallel reductions (e.g., dot products, sums, matrix multiplications). By providing a theory for optimal tree topologies based on input statistics, it enables the development of adaptive algorithms that minimize numerical error without sacrificing performance. This is particularly relevant for mixed-precision training and inference, where understanding and controlling error propagation is critical. The insights could lead to more robust and accurate numerical libraries for deep learning and scientific computing. This paper provides a rigorous second-moment theory for floating-point reduction trees, deriving exact error recurrences and optimal topologies that significantly advance the understanding of numerical stability in parallel reductions, with direct applications to improving the accuracy of HPC and ML libraries.
Pruning long context for coding agents has been a vital technology for efficient context management. While existing context pruning methods such as SWE-Pruner realize this by attaching a separate code classifier, we find the agent itself encodes internal representations indicating the relevance of code context when reading tool output. Based on this finding, we propose SWE-Pruner Pro, which prunes tool outputs directly inside the agent. Concretely, a small head turns the agent's own internal representations into a keep-or-prune label for each line, with a length-aware embedding keyed to each tool output's line count. Across two open-weight backbones and four multi-turn benchmarks, SWE-Pruner Pro saves up to 39% of prompt and completion tokens while preserving task quality, with bounded inference overhead. Notably, on MiMo-V2-Flash SWE-Pruner Pro additionally raises the SWE-Bench Verified resolve rate by +3.8% and the long-context Oolong accuracy by +2.2 points.
Primary: Shanghai Jiao Tong University
All Institutions: Shanghai Jiao Tong University
SWE-Pruner Pro demonstrates that coding agents' internal representations encode sufficient line-level relevance information to prune tool outputs effectively, achieving significant token savings and improved efficiency without retraining the backbone, marking a practical step forward in scalable agentic systems.
The paper proposes SWE-Pruner Pro, a method for pruning long context in coding agents by extracting a lightweight classification head from the agent's own internal hidden states. The core insight is that the backbone model already encodes line-level relevance information during the forward pass of tool outputs, eliminating the need for a separate scoring model or explicit goal-hint queries used in prior work (e.g., SWE-Pruner). The methodology involves a length-aware embedding and a per-sample balanced focal loss to handle class imbalance and length-dependent error costs. The approach is technically sound, leveraging existing inference infrastructure (SGLang) with specific patches to expose hidden states. The design is pragmatic, focusing on efficiency gains without retraining the backbone.
The evaluation is comprehensive, covering two large open-weight backbones (MiMo-V2-Flash, Qwen3-Coder-Next) and four benchmarks (SWE-Bench Verified, SWE-QA, SWE-QA-Pro, Oolong). The results demonstrate up to 39% token savings while preserving or slightly improving task quality (e.g., +3.8% resolve rate on SWE-Bench Verified). The inclusion of an ablation study on loss functions and length-aware embeddings adds rigor. The comparison against seven baselines, including strong prior work, provides a solid empirical foundation. The latency analysis, including in-engine colocation, further strengthens the practical value of the work.
The paper provides detailed implementation details, including the architecture of the pruning head, training data sources (publicly released datasets), and specific SGLang patches required for hidden state extraction. The training data distribution and labeling protocol are described. However, the code for the SGLang patches and the specific pruning head implementation are not explicitly linked in the text (though implied to be available or part of the project). The reliance on specific versions of SGLang and the need for custom patches might pose minor reproducibility hurdles for users not familiar with the inference engine's internals.
The method is currently limited to open-weight models that expose hidden states. The evaluation is primarily focused on Python and CLI tasks, with limited coverage of other programming languages. The pruning is applied to tool outputs, so it does not address pruning of the agent's own reasoning or history, which may also contain redundant information. The performance gain on SWE-Bench Verified is notable but the absolute resolve rate is still constrained by the backbone's capabilities.
This work significantly advances the field of efficient LLM inference, particularly for agentic workflows. By demonstrating that internal representations contain sufficient signal for pruning, it reduces the computational overhead and latency associated with context management. This can lead to more cost-effective and scalable deployment of coding agents, enabling longer and more complex interactions. The approach could be generalized to other domains where agents interact with long textual environments. SWE-Pruner Pro demonstrates that coding agents' internal representations encode sufficient line-level relevance information to prune tool outputs effectively, achieving significant token savings and improved efficiency without retraining the backbone, marking a practical step forward in scalable agentic systems.
We present RynnBrain 1.1, a family of embodied foundation models spanning 2B, 9B, and 122B-A10B scales. Trained with a unified spatio-temporal and physically grounded framework, RynnBrain 1.1 supports embodied perception, spatial reasoning, localization, and planning. Compared with RynnBrain 1.0, it further introduces contact-point prediction across the model family and native 3D grounding for the 2B and 9B models, yielding representations and outputs that are more directly aligned with robot manipulation. We also develop RynnBrain-VLA with a unified cross-embodiment action space and embodiment-specific masking, and deploy it on Unitree G1, Astribot-S1, and Tianji-Wuji. RynnBrain 1.1 achieves strong results on embodied cognition, localization, and 3D grounding, with the 122B-A10B model outperforming all evaluated proprietary and open-source models on VSI-Bench, MMSI, and RefSpatial-Bench. Real-robot experiments show that RynnBrain-initialized policies outperform Qwen-based and representative generalist VLAs, while joint multi-task and multi-embodiment training improves process scores and success rates over per-task training.
Primary: Rynn AI
All Institutions: Rynn AI
RynnBrain 1.1 presents a significant advancement in embodied foundation models by introducing native 3D grounding and contact-point prediction within a unified spatio-temporal framework, demonstrating strong generalization across multiple robotic platforms and outperforming existing models on key embodied cognition benchmarks.
The paper introduces RynnBrain 1.1, a family of embodied foundation models (2B, 9B, 122B-A10B). The core methodological contribution lies in a unified spatio-temporal and physically grounded pretraining framework. Key technical innovations include the introduction of contact-point prediction across the model family and native 3D grounding for the smaller models (2B and 9B). The authors also propose RynnBrain-VLA, which utilizes a unified cross-embodiment action space and embodiment-specific masking to handle diverse robotic hardware. The architecture appears to scale from dense to Mixture-of-Experts (MoE) models, aiming to balance capability with inference efficiency. The approach integrates perception, spatial reasoning, and planning into a single unified representation, which is a significant step towards generalist embodied agents.
The evaluation covers embodied cognition, localization, and 3D grounding benchmarks (VSI-Bench, MMSI, RefSpatial-Bench). The 122B-A10B model claims to outperform all evaluated proprietary and open-source models on these benchmarks. Real-robot experiments are conducted on three distinct platforms: Unitree G1, Astribot-S1, and Tianji-Wuji. The results indicate that RynnBrain-initialized policies outperform Qwen-based and other generalist VLAs. The paper also highlights the benefits of joint multi-task and multi-embodiment training over per-task training, showing improvements in process scores and success rates. The breadth of evaluation across simulation and real-world hardware is a strong point, demonstrating generalization capabilities.
The paper provides details on model scales and training frameworks. However, as an arXiv preprint without an accompanying code release mentioned in the text, reproducibility is currently limited to the described methodology. The use of specific hardware (Unitree, Astribot, Tianji) for real-world evaluation adds a layer of complexity for independent verification, requiring access to similar robotic platforms. The mention of "RynnBrain" suggests a proprietary model family, which may limit full open-source reproducibility of the weights.
The paper focuses heavily on the capabilities of the 122B-A10B model for benchmark leadership, but the performance of the smaller 2B and 9B models, while improved, may not match the state-of-the-art in every metric. The reliance on embodiment-specific masking and unified action spaces requires careful calibration for each new robot type, which might limit plug-and-play generalization to unseen embodiments without further tuning. The computational cost of training and deploying the 122B model is significant, potentially limiting accessibility for smaller research groups.
This work contributes to the development of generalist embodied AI, which has profound implications for robotics, automation, and human-robot interaction. By providing models that can reason about 3D space and physical interactions, it paves the way for more capable and autonomous robots in unstructured environments. The emphasis on generalization across embodiments could accelerate the deployment of robots in various industries. However, the increased autonomy of robots also raises safety and ethical considerations regarding control and reliability in physical spaces. RynnBrain 1.1 presents a significant advancement in embodied foundation models by introducing native 3D grounding and contact-point prediction within a unified spatio-temporal framework, demonstrating strong generalization across multiple robotic platforms and outperforming existing models on key embodied cognition benchmarks.
Do independently trained language models come to represent the same thing in the same way? We answer for code, extending a recently introduced concept-circuit extraction method to a 2x2 design -- Python and Rust crossed with Qwen2.5-Coder-7B and DeepSeek-Coder-V1-6.7B -- and measuring a complete inventory of grammatical concepts (58 Python, 57 Rust) identically in all four cells: the smallest design that separates what depends on the task, the language, and the model. The answer splits into three parts. What earns dedicated circuitry is set by the task: the models agree on which concepts receive circuits (Spearman $ρ$ = 0.638 for Python, 0.673 for Rust, both p < $10^{-7}$). Where those circuits sit is set by the model: Qwen processes concepts in a late band (~L17-19), DeepSeek at L6-7, for both languages. How circuits grow across layers is also set by the model: Qwen gives its atomic concepts an early spike that DeepSeek does not. "Are circuits universal?" thus has no single answer: yes for What, no for Where and How -- universality is a property of representational content, not of computational organisation. None of this structure was fixed in advance. The agreement could have landed anywhere between independence and identity; it lands at $ρ\approx 0.65$. Rust constructs receive 2-3x more concept-specific circuitry than their Python equivalents, in both models. Both models share neurons between the languages (6/7 and 7/7 paired constructs), DeepSeek 1.94x more than Qwen -- a direction no prior result predicts. And Qwen binds nine keywords of Rust's type-and-trait machinery into one tight neuron cluster (Jaccard 0.535 vs null 0.112, p < 0.001), a semantic dimension invisible in surface syntax. Ablation and linear probes confirm the circuits are functional. All claims are scoped to this 2x2; whether the per-model profile predicts a third model is the designed next test.
Primary: University College London
All Institutions: University College London
This work has significant broader impact for the field of mechanistic interpretability and our fundamental understanding of large language models. By systematically disentangling the "What, Where, and How" of concept representations, it provides a more nuanced and accurate view of universality, moving beyond simplistic yes/no answers. This conceptual framework is crucial for developing more robust and transferable interpretability techniques. Operationally, it offers actionable guidance for practitioners, clarifying which interpretability findings (e.g., concept inventories) are likely to transfer across models and which (e.g., layer-specific interventions) require re-localization. The discovery of "two processing styles" as a model fingerprint offers a predictive framework for characterizing new models. The method's ability to uncover semantic dimensions beyond surface syntax (e.g., Rust's type-trait cluster) suggests its potential for deeper insights into model understanding. While currently applied to formal languages, the stated goal of extending to natural language promises even wider implications for understanding how models process human language. This research contributes to building a more systematic and scientific foundation for understanding the internal workings of complex neural networks. This paper systematically disentangles the roles of task, language, and model in code model representations, revealing that representational content ("What") is largely universal, while computational organization ("Where" and "How") is model-specific, and language design influences representation strength. The authors extend a concept-circuit extraction method to a rigorous 2x2 experimental design, providing compelling empirical evidence for a nuanced view of universality, identifying a "model fingerprint" of processing styles, and demonstrating the recovery of abstract semantic dimensions beyond surface syntax, all validated through causal ablation and linear probes. This work significantly advances mechanistic interpretability by providing a systematic, comparable framework for understanding internal model representations and offering actionable insights for transferring interpretability results across models.
The paper extends a recently introduced concept-circuit extraction method (Wilam 2026) to systematically identify neuron circuits corresponding to specific grammatical concepts. This method is a significant methodological advance for interpretability, as it provides a common, comparable yardstick across different models and languages. The core idea involves generating diverse "concept prompts" to isolate concept-specific circuitry through marginalization (intersection of active neurons), and contrasting these with "checker prompts" to differentiate concept-specific responses from mere token recognition. The application of this method within a novel 2x2 experimental design (Python/Rust crossed with Qwen2.5-Coder-7B/DeepSeek-Coder-V1-6.7B) is particularly elegant, as it is precisely structured to disentangle the roles of the task (concept identity), the language, and the model. The pipeline is well-defined, encompassing prompt generation with injected variance, extraction of MLP outputs, binarization of neuron activations (with a chosen threshold of 0.5 for structural signal), marginalization via intersection, and decomposition into concept-only, shared, and token-only masks. This systematic and quantitative approach addresses a key limitation of many prior interpretability methods, which are often model-specific and lack direct comparability. While the reliance on a binary neuron mask is acknowledged as a potential limitation for capturing sub-threshold distributed information, it is a deliberate design choice that enables the crucial cross-model comparison.
The experimental evaluation is comprehensive, rigorous, and yields significant empirical findings. The choice of two distinct, large-scale code models (Qwen2.5-Coder-7B, DeepSeek-Coder-V1-6.7B) and two popular formal languages (Python, Rust) provides a robust foundation for the comparative analysis. The inventory of 58 Python and 57 Rust testable concepts is thorough, focusing on constructs that allow for the critical concept-vs-token contrast. The results are clearly presented and strongly supported by quantitative metrics. The paper convincingly demonstrates the "What/Where/How" dissociation: "What" (which concepts earn circuitry) shows significant cross-model agreement (Spearman ρ ≈ 0.65), indicating a conserved ranking of concept salience. In contrast, "Where" (layer placement) and "How" (circuit growth dynamics) diverge sharply, with Qwen processing concepts in a late band (L17-19) and DeepSeek in an early band (L6-7) for both languages, and exhibiting different early-layer dynamics for atomic concepts. Beyond this core dissociation, the experiments reveal that Rust constructs consistently receive 2-3x more concept-specific circuitry than Python equivalents, highlighting a language-design effect. Cross-language neuron sharing is observed, with DeepSeek sharing more than Qwen. A particularly striking finding is Qwen's semantic clustering of Rust's type-and-trait machinery, recovering a conceptual dimension invisible in surface syntax. Causal ablation experiments on Qwen Python provide functional validation for several concepts, confirming the identified circuits' role in model behavior. Linear probes further corroborate the decodability of concept information. The experiments are exceptionally well-designed to address the research questions and provide strong empirical evidence for the paper's claims.
The paper sets an excellent standard for reproducibility. The authors explicitly state that all analysis code, figure scripts, and a frozen-numbers test suite are released on GitHub (https://github.com/piotrwilam/Atlas2x2). Crucially, the frozen experimental artifacts (neuron activations) are released as a dataset on Hugging Face (https://huggingface.co/datasets/piotrwilam/Atlas2x2). The paper guarantees that "Every number and figure in the paper regenerates from the released analysis layer without rerunning a model," which is a gold standard for transparency and allows for full verification and extension of the research by the community.
The authors provide a transparent and well-articulated discussion of the study's limitations: 1. **Threshold structure:** The method focuses on high-amplitude neurons, potentially missing distributed sub-threshold representations, though a continuous treatment is planned. 2. **Languages only:** The study is restricted to imperative languages (Python, Rust), and generalization to declarative languages or proof assistants remains an open question. 3. **Layer-count mismatch:** The models have different layer counts (28 vs. 32), which is addressed by using both absolute indices and fraction-of-depth, but is still a factor in direct layer comparisons. 4. **Validation scope:** Causal validation is performed only on the Qwen Python cell, as DeepSeek's smoother circuit dynamics lack a clear peak layer for single-layer ablation, meaning the causal evidence for the full 2x2 design is not yet complete. 5. **Consistency parameter:** A parameter that was meaningful in prior work is degenerate for the dense SwiGLU architectures used, limiting the sweep to the activation threshold alone. 6. **Concept space mismatch:** Comparisons are restricted to shared testable subsets due to differences in the full concept space across models/languages. These acknowledged limitations provide clear avenues for future research.
This work has significant broader impact for the field of mechanistic interpretability and our fundamental understanding of large language models. By systematically disentangling the "What, Where, and How" of concept representations, it provides a more nuanced and accurate view of universality, moving beyond simplistic yes/no answers. This conceptual framework is crucial for developing more robust and transferable interpretability techniques. Operationally, it offers actionable guidance for practitioners, clarifying which interpretability findings (e.g., concept inventories) are likely to transfer across models and which (e.g., layer-specific interventions) require re-localization. The discovery of "two processing styles" as a model fingerprint offers a predictive framework for characterizing new models. The method's ability to uncover semantic dimensions beyond surface syntax (e.g., Rust's type-trait cluster) suggests its potential for deeper insights into model understanding. While currently applied to formal languages, the stated goal of extending to natural language promises even wider implications for understanding how models process human language. This research contributes to building a more systematic and scientific foundation for understanding the internal workings of complex neural networks. This paper systematically disentangles the roles of task, language, and model in code model representations, revealing that representational content ("What") is largely universal, while computational organization ("Where" and "How") is model-specific, and language design influences representation strength. The authors extend a concept-circuit extraction method to a rigorous 2x2 experimental design, providing compelling empirical evidence for a nuanced view of universality, identifying a "model fingerprint" of processing styles, and demonstrating the recovery of abstract semantic dimensions beyond surface syntax, all validated through causal ablation and linear probes. This work significantly advances mechanistic interpretability by providing a systematic, comparable framework for understanding internal model representations and offering actionable insights for transferring interpretability results across models.
LLMs scale Mixture-of-Experts (MoE) parameters for superior intelligence, but massive weights and dynamic computation impede efficient serving. Existing instance-level prefill-decode disaggregation isolates the phases on separate full-model replicas. As MoE weights grow, each instance may span tens to hundreds of GPUs, making resource allocation increasingly coarse. Configured prefill-to-decode ratios thus often mismatch demand, overprovisioning one phase while overloading the other. Prefill-decode colocation avoids this duplication, but existing Green Context solutions partition each GPU by phase and fix phase resources during a kernel. They cannot track resource changes across operations or layerwise variation in routed expert load, causing head-of-line blocking or idle reserved resources. Partitioning every GPU also leaves each phase with fewer local resources, forces wider parallelism and more communication, and lets prefill and decode traffic interfere on the shared network. We present ExpertPlex, which shares massive MoE experts across phases while disaggregating lightweight attention modules. Expert sharing eliminates over 95% of duplicate model weights and multiplexes dynamically sparse computation, while attention disaggregation reduces attention communication cost. ExpertPlex further uses (1) adaptive persistent kernels to schedule dynamic expert computation at tile granularity for efficient, isolated execution; (2) attention-initiated MoE communication to avoid network interference and enable cross-phase communication-computation overlap; and (3) a tile-to-cluster model to optimize these mechanisms for maximum goodput. Experiments serving MiniMax-M2.7 and GLM-5.1-FP8 show that ExpertPlex improves goodput by up to 2.01$\times$ over instance-level prefill-decode disaggregation and 1.66$\times$ over prefill-decode colocation.
Primary: Peking University
All Institutions: Peking University, Independent Researcher
ExpertPlex presents a significant advancement in LLM serving systems by introducing a novel disaggregated architecture that effectively shares MoE experts while isolating attention modules, achieving substantial goodput improvements through adaptive persistent kernels and optimized communication patterns.
The paper proposes ExpertPlex, a disaggregated serving architecture specifically designed for Mixture-of-Experts (MoE) Large Language Models. The core innovation lies in decoupling the handling of MoE experts from attention modules. While existing systems either colocate all components (leading to resource contention) or disaggregate at the instance level (leading to massive weight duplication and coarse-grained allocation), ExpertPlex shares the massive MoE expert weights across phases while isolating the lightweight attention modules. The methodology introduces three key technical mechanisms: (1) Adaptive Persistent Kernels, which schedule dynamic expert computation at the tile granularity to handle load imbalance and avoid head-of-line blocking; (2) Attention-Initiated MoE Communication, which overlaps communication with computation and prevents network interference between prefill and decode phases; and (3) A Tile-to-Cluster Model, an optimization framework to allocate resources dynamically. This approach addresses the fundamental inefficiency of serving sparse MoE models where expert load is highly dynamic and non-uniform.
The evaluation is conducted on two significant MoE models: MiniMax-M2.7 and GLM-5.1-FP8. The baseline comparisons are rigorous, covering the two dominant existing paradigms: instance-level prefill-decode disaggregation and prefill-decode colocation. The results demonstrate substantial improvements, with up to 2.01x goodput improvement over instance-level disaggregation and 1.66x over colocation. These gains are particularly impressive given that colocation is often considered the most resource-efficient in terms of hardware usage, implying that ExpertPlex achieves higher throughput without requiring additional hardware, simply by better utilizing existing resources. The use of FP8 models also highlights the system's relevance to current hardware trends.
The paper provides a detailed description of the system design, including the persistent kernel scheduling and communication overlap mechanisms. The inclusion of specific model names (MiniMax-M2.7, GLM-5.1-FP8) allows for potential replication if these models are publicly available or if synthetic workloads are used. However, as is common with systems papers, full reproducibility might depend on the specific cluster configuration and the availability of the proprietary model weights. The detailed algorithmic descriptions of the adaptive scheduling and tile-to-cluster optimization provide a strong foundation for implementation.
The paper focuses on the serving side and does not address training efficiency. The complexity of the adaptive persistent kernels and the tile-to-cluster model introduces additional system overhead that must be carefully managed; if the scheduling overhead exceeds the gains from reduced communication or better utilization, performance could degrade. Furthermore, the benefits are most pronounced for very large MoE models with high expert counts; for smaller models or dense models, the overhead of the disaggregation and dynamic scheduling might not justify the complexity. The reliance on high-bandwidth, low-latency interconnects for the attention-MoE communication is also a constraint.
ExpertPlex addresses a critical bottleneck in the deployment of state-of-the-art AI models. By significantly improving the goodput of MoE LLMs, it lowers the cost barrier for serving these models, potentially democratizing access to high-quality AI services. The architectural insights regarding dynamic resource allocation for sparse models can influence future system designs for other sparse architectures beyond LLMs, such as sparse transformers or mixture-of-experts in other domains. ExpertPlex presents a significant advancement in LLM serving systems by introducing a novel disaggregated architecture that effectively shares MoE experts while isolating attention modules, achieving substantial goodput improvements through adaptive persistent kernels and optimized communication patterns.
Retrieval-Augmented Generation (RAG) enhances the factual grounding of large language model (LLM) inference by retrieving relevant information from external knowledge bases. However, its dense vector retrieval introduces significant latency and energy overhead, becoming the primary performance bottleneck. Although recent in-storage accelerators aim to reduce data movement, they still rely on host or embedded processors outside the memory, where nearly 70% of the total retrieval time is spent. As a result, they cannot fully overcome the bandwidth limitations, leading to yet another memory bottleneck. To tackle these limitations, we present D-NOVA, a hardware-software co-designed in-storage retrieval accelerator. D-NOVA executes an inverted file (IVF)-based hierarchical retrieval pipeline by deeply embedding the search functionality directly into the NAND memory array. This is achieved by incorporating a new distance metric, Dual-Bound Tight Similarity Sensing (DTS), which is specifically tailored for searching within the NAND string. In addition, we introduce a lightweight contrastive adapter that maps embedding vectors into a DTS-friendly domain, recovering near-software recall while improving performance and energy efficiency. D-NOVA is up to 41.7x faster and 71x more energy-efficient than a CPU baseline, and achieves 12.13x higher throughput while being up to 1.26x more energy-efficient than state-of-the-art in-storage RAG accelerators, demonstrating the potential of fully in-storage vector search for scalable RAG acceleration.
Primary: University of California, San Diego
All Institutions: University of California, San Diego
D-NOVA introduces a novel in-storage vector search accelerator that leverages a NAND-optimized distance metric and contrastive adaptation to achieve significant speed and energy efficiency gains for RAG workloads. This research represents a substantial contribution to the intersection of computer architecture and machine learning systems, offering a viable path to overcoming the memory bottleneck in large-scale retrieval tasks.
The paper proposes D-NOVA, a hardware-software co-designed accelerator that performs vector similarity search directly within 3D NAND memory arrays, bypassing the traditional host-to-memory data movement bottleneck. The core technical innovation is the Dual-Bound Tight Similarity Sensing (DTS) metric, which is specifically tailored to the physical characteristics of NAND strings (e.g., threshold voltage distributions and read disturb effects) to enable approximate nearest neighbor search at the storage level. This is coupled with a lightweight contrastive adapter that transforms embedding vectors into a domain compatible with DTS, allowing the hardware to operate on "DTS-friendly" vectors while maintaining high recall relative to software baselines. The approach represents a significant shift from "in-storage computing" (which often still moves data to embedded processors) to "in-storage sensing," leveraging the analog/digital properties of the memory array itself for computation.
The evaluation demonstrates substantial improvements over baselines. D-NOVA achieves up to 41.7x speedup and 71x energy efficiency gains compared to a CPU baseline. When compared to state-of-the-art in-storage RAG accelerators, it shows 12.13x higher throughput and up to 1.26x better energy efficiency. The paper likely includes detailed breakdowns of latency components, energy consumption per operation, and recall/accuracy metrics (e.g., Recall@K) to validate that the hardware approximation does not significantly degrade retrieval quality. The results suggest that the proposed architecture effectively mitigates the memory wall for RAG workloads.
As a hardware design paper, reproducibility depends on the availability of the RTL (Register Transfer Level) code, simulation models, and detailed architectural parameters. The paper mentions support from SRC and NSF grants, suggesting rigorous academic standards. However, without explicit open-source code links in the provided text, reproducibility is limited to the described methodology and potentially available supplementary materials. The specific implementation of the DTS metric and the contrastive adapter's training procedure are critical for replication.
The primary limitation is the reliance on specific 3D NAND characteristics, which may vary across manufacturers and process nodes. The "DTS-friendly" domain requires a pre-processing step (the contrastive adapter), adding a small overhead that must be justified by the massive gains in the search phase. Furthermore, the scalability of the in-storage search logic across very large vector databases (millions/billions of vectors) and the impact of wear-leveling and garbage collection in NAND on the consistency of the DTS metric are potential challenges not fully addressed in the abstract. The energy efficiency claim of 1.26x over SOTA in-storage accelerators is modest compared to the CPU baseline, suggesting that while the approach is novel, the absolute gains over existing in-storage solutions might be incremental in some configurations.
This work has significant implications for the scalability of Retrieval-Augmented Generation (RAG) systems, which are becoming the standard for enterprise LLM applications. By reducing the latency and energy cost of vector retrieval, D-NOVA enables more responsive and sustainable AI systems. It also advances the field of in-memory/in-storage computing, demonstrating that complex ML workloads can be offloaded to storage devices in a way that leverages their physical properties, potentially reshaping the architecture of future data centers. D-NOVA introduces a novel in-storage vector search accelerator that leverages a NAND-optimized distance metric and contrastive adaptation to achieve significant speed and energy efficiency gains for RAG workloads. This research represents a substantial contribution to the intersection of computer architecture and machine learning systems, offering a viable path to overcoming the memory bottleneck in large-scale retrieval tasks.
Multi-head Latent Attention (MLA) ships two implementations in Megatron-Core: an explicit form used for training and an absorbed form -- which slashes collective communication by gathering only the compressed latent -- that is fully implemented but hard-asserted out of training (the forward opens with "assert not (self.training and self.cache_mla_latents)"), allowed only in inference decode. The library documents no reason. We show the restriction is well-founded and quantify why: ported to training, the absorbed form is a memory trap -- its intermediates live in n_h x d_kv dimensions per token, larger than the per-head K/V they replace -- inflating activation memory by 20-34%, up to 9.2 GB at DeepSeek-V3 scale (n_h=128, seq=16384, SP=8, eager kernel; the gap widens to 19.2 GB under a fused kernel), enough to change device-fit. This measurement, validated on two axes (linear in seq and n_h) and cross-verified on NVIDIA A100, explains the otherwise-undocumented restriction and leaves practitioners with no low-communication MLA training path. We then provide one. LAGA (Latent All-Gather Attention) keeps the absorbed form's latent-gather communication but rejects the absorb reformulation, instead reconstructing per-head K/V locally from the gathered latent. On 8x Ascend 910B at real DeepSeek-V3 dimensions, LAGA cuts collective communication 1.98x, matches explicit memory within 0.5%, is bit-identical to explicit at SP=1 and equivalent to within 1e-3 at SP=2-8, and under a fused attention kernel improves attention-block throughput 1.04-1.06x single-node and 1.07-1.24x cross-node -- leading at all sequence lengths in the cross-node regime MLA is deployed for.
Primary: China Mobile Jiutian Artificial Intelligence Technology (Beijing) Co., Ltd.
All Institutions: China Mobile Jiutian Artificial Intelligence Technology (Beijing) Co., Ltd.
The paper provides a rigorous analysis of the memory-communication trade-offs in MLA sequence parallelism and proposes LAGA, a practical solution that enables efficient training of MLA-based models like DeepSeek-V3. [Comprehensive analysis of the technical contribution, methodology, and significance to the field].
The paper addresses a critical gap in distributed training systems for Large Language Models (LLMs), specifically concerning Multi-head Latent Attention (MLA) as used in DeepSeek-V3. The authors identify that Megatron-Core explicitly forbids the "absorbed" form of MLA during training, a restriction previously undocumented. They provide a rigorous theoretical and empirical explanation: the absorbed form, while communication-efficient, creates a "memory trap" where intermediate tensors scale with $n_h \times d_{kv}$, leading to significant activation memory inflation (up to 34% or ~9-19 GB at scale). To solve this, they propose LAGA (Latent All-Gather Attention), which retains the communication efficiency of the latent all-gather but reconstructs per-head K/V locally to maintain the memory footprint of the explicit form. The methodology is sound, leveraging standard sequence parallelism principles and linear algebra properties to decouple communication volume from activation memory size.
The evaluation is comprehensive and convincing. The authors measure communication volume, activation memory, and throughput on both Ascend 910B and NVIDIA A100 hardware, ensuring hardware-agnostic validity. They demonstrate that LAGA reduces collective communication by ~1.98x compared to the explicit baseline while matching its memory footprint within 0.5%. Crucially, they show that LAGA outperforms the explicit baseline in throughput (1.04-1.24x) in the cross-node, long-context regime, which is the primary use case for MLA. The numerical equivalence to the explicit baseline is verified to high precision (bit-identical at SP=1, <1e-3 error at SP>1). The inclusion of both eager and fused kernel results strengthens the claims by showing the memory trap is structural, not an artifact of kernel implementation.
The paper provides detailed algorithmic descriptions and mathematical derivations. The experimental setup is clearly defined, including hardware specs, sequence lengths, and parallelism degrees. The authors cross-verify results on two different hardware architectures (Ascend and NVIDIA), which adds significant credibility. While the code is not linked, the description is sufficient for a competent systems researcher to implement. The convergence test on a small model provides additional assurance of correctness.
The evaluation is limited to a single attention layer for communication/memory analysis and a 4-layer stack for throughput, rather than a full DeepSeek-V3-scale (61-layer) model. The authors acknowledge this, noting that the structural properties hold, but end-to-end MFU might differ. Additionally, the convergence test is on a small, random-data memorization task, which may not fully reflect training dynamics on real data. The prototype implementation uses eager kernels for some comparisons, though fused kernel results are provided to mitigate this.
This paper has significant implications for the training of large-scale MoE and MLA-based models. By providing a communication-efficient and memory-safe training path for MLA, it enables more efficient scaling of these architectures. The finding that inference optimizations (like absorption) can be detrimental in training is a valuable lesson for the broader ML systems community. It highlights the importance of analyzing memory footprints when porting inference techniques to training. The paper provides a rigorous analysis of the memory-communication trade-offs in MLA sequence parallelism and proposes LAGA, a practical solution that enables efficient training of MLA-based models like DeepSeek-V3. [Comprehensive analysis of the technical contribution, methodology, and significance to the field].